Modeling marine cargo traffic to identify countries in Africa with greatest risk of invasion by Anopheles stephensi – Scientific Reports
With human movement and globalization, invasive container breeding vectors responsible for dengue, Zika, chikungunya and now malaria, with An. stephensi, are being introduced and establishing populations in new locations. They are bringing with them the threat of increasing or novel cases of vector-borne diseases to new locations where health systems may not be prepared.
Anopheles stephensi was first detected on the African continent in Djibouti in 2012 and has since been confirmed in Ethiopia, Somalia, and Sudan. Unlike most malaria vectors, An. stephensi is often found in artificial containers and in urban settings. This unique ecology combined with its initial detection in seaports in Djibouti, Somalia, and Sudan has led scientists to believe that the movement of this vector is likely facilitated through maritime trade.
By modeling inter- and intra-continental maritime connectivity in Africa we identified countries with higher likelihood of An. stephensi introduction if facilitated through maritime movement and ranked them based on this data. Anopheles stephensi was not detected in Africa (Djibouti) until 2012. To determine whether historical maritime data would have identified the first sites of introduction, 2011 maritime data were analyzed to determine whether the sites with confirmed An. stephensi would rank highly in connectivity to An. stephensi endemic countries. Using 2011 data on maritime connectivity alone, Djibouti and Sudan were identified as the top two countries at risk of An. stephensi introduction if it is facilitated by marine cargo shipments. In 2021, these are two of the three African coastal nations where An. stephensi is confirmed to be established.
When 2011 maritime data were combined with the HSI for An. stephensi establishment, the top five countries remain the same as with maritime data alone: Sudan, Djibouti, Egypt, Kenya and Tanzania, in that order. The maritime data show likelihood of introduction and HSI shows likelihood of establishment. When combined, the analyses show a likelihood of being able to establish and survive once introduced. Interestingly, the results of the combined analyses align with the detection data being reported in the Horn of Africa. The 2011 maritime data reinforces the validity of the model as it points to Sudan and Djibouti, where An. stephensi established in the following years. Similarly, the HSI data for Ethiopia has aligned closely with detections of the species to date15. Interestingly, around this time of initial detection in Djibouti, Djibouti City port underwent development and organizational change. The government of Djibouti took back administrative control of the port as early as 201230.
Following this method, maritime trade data from 2020 could point to countries at risk of An. stephensi introduction from endemic countries as well as from the coastal African countries with newly introduced populations. Here we provide a prioritization list and heat map of countries for the early detection, rapid response, and targeted surveillance of An. stephensi in Africa based on this data and the HSI (Fig. 4). Further invasion of An. stephensi on the African continent has the potential to reverse progress made on malaria control in the last century. Anopheles stephensi thrives in urban settings and in containers, in contrast to the rural settings and natural habitats where most Anopheles spp. are found20. The situation in Djibouti may be a harbinger for what is to come if immediate surveillance and control strategies are not initiated18.
Maritime data from 2020, with Djibouti and Sudan considered as potential source populations for intracontinental introduction of An. stephensi, indicate the top five countries at risk for maritime introduction are Egypt, Kenya, Mauritius, Tanzania, and Morocco, suggesting that targeted larval surveillance in these countries near seaports may provide a better understanding of whether there are maritime introductions. When the data from 2020 data is combined with HSI for An. stephensi, the top five countries are instead Egypt, Kenya, Tanzania, Morocco, and Libya. Interestingly, historical reports of An. stephensi in Egypt exist; however, following further identification these specimens were determined to be An. ainshamsi31. With several suitable habitats both along the coast and inland of Egypt, revisiting surveillance efforts there would provide insight into how countries that are highly connected to An. stephensi locations through maritime traffic may experience introductions.
Further field validation of this prioritization list is necessary, because it is possible that An. stephensi is being introduced through other transportation routes, such as dry ports or airports32, or may even be dispersed through wind facilitation33. However, countries highlighted here with high levels of connectivity to known An. stephensi locations should be considered seriously at risk and surveillance urgently established to determine whether An. stephensi introduction has already occurred or to enable early detection. Primary vector surveillance for both Ae. aegypti and An. stephensi are through larval surveys, and the two mosquitoes are commonly detected in the same breeding habitats. It could therefore be beneficial to coordinate with existing Aedes surveillance efforts to be able to simultaneously gather data on medically relevant Aedes vectors while seeking to determine whether An. stephensi is present. Similarly, in locations with known An. stephensi and not well established Aedes programs, coordinating surveillance efforts provides an opportunity to conduct malaria and arboviral surveillance by container breeding mosquitoes simultaneously.
Efforts to map pinch points or key points of introduction based on the movement of goods and populations could provide high specificity for targeted surveillance and control efforts. For example, participatory mapping or population mobility data collection methods, such as those used to determine routes of human movement for malaria elimination, may simultaneously provide information on where targeted An. stephensi surveillance efforts should focus. Several methods have been proposed in the literature for modeling human movement and one in particular, PopCAB, which is often used for communicable diseases, combined quantitative and qualitative data with geospatial information to identify points of control34.
Data on invasive mosquito species has shown that introduction events are rarely a one-time occurrence. Population genetics data on Aedes species indicate that reintroductions are very common and can facilitate the movement of genes between geographically distinct populations, raising the potential for introduction of insecticide resistance, thermotolerance, and other phenotypic and even behavioral traits which may be facilitated by gene flow and introgression35. Djibouti, Sudan, Somalia, and Ethiopia, countries with established invasive populations of An. stephensi, should continue to monitor invasive populations and points of introduction to control and limit further expansion and adaptation of An. stephensi. Work by Carter et al. has shown that An. stephensi populations in Ethiopia in the north and central regions can be differentiated genetically, potentially indicating that these populations are a result of more than one introduction into Ethiopia from South Asia, further emphasizing the potential role of anthropogenic movement on the introduction of the species17.
One major limitation of this work is that Somalia is the third coastal nation where An. stephensi has been confirmed; however, marine traffic data were not available for Somalia so it could not be included in this analysis. The potential impact of Somalia on maritime trade is unknown and it should not be excluded as a potential source population. Additionally, this model does not account for the possibility of other countries with An. stephensi populations that have not been detected yet. As new data on An. stephensi expansion becomes available, more countries will be at higher risk. Other countries with An. stephensi populations, such as Iran, Myanmar, and Iraq, constitute lower relative percentages of trade with these countries so were not included in the analysis. However, genetic similarities were noted from An. stephensi in Pakistan, so this nation was included10.
Due to the nature of maritime traffic, inland countries were also not included in this prioritization ranking. Countries which are inland but share borders with high-risk countries according to the LASTIMI index should also be considered with high priority. For example, the ranking from 2011 highlights Sudan and Djibouti, both which border Ethiopia, and efforts to examine key land transportation routes between bordering nations where humans and goods travel may provide additional insight into the expansion routes of this invasive species.
In Ethiopia, An. stephensi was detected in 2016. It has largely been detected along major transportation routes although further data is needed to understand the association between movement and An. stephensi introductions and expansion since most sampling sites have also been located along transport routes. Importantly, Ethiopia relies heavily on the ports of Djibouti and Somalia for maritime imports and exports. Surveillance efforts have revealed that the species is also frequently associated with livestock shelters and An. stephensi are frequently found with livestock bloodmeals15. Interestingly, the original detection of An. stephensi was found in a livestock quarantine station in the port of Djibouti. Additionally, livestock constitutes one of the largest exports of maritime trade from this region. For countries with high maritime connectivity to An. stephensi locations, surveillance efforts near seaports, in particular those with livestock trade, may be targeted locations for countries without confirmed An. stephensi to begin larval surveillance.
As Ae. aegypti and Culex coronator were detected in tires or Ae. albopictus through tire and bamboo (Dracaena sanderiana) trade, An. stephensi could be carried through maritime trade of a specific good36,37,38. Future examination of the movement of specific goods would be beneficial in interpreting potential An. stephensi invasion pathways. Additionally, the various types of vessels used to transport certain cargo such as container, bulk, and livestock ships could affect An. stephensi survivability during transit. Sugar and grain are often shipped in bulk or break bulk vessels which store cargo in large unpackaged containers. Container ships transport products stored in containers sized for land transportation via trucks and carry goods such as tires. Livestock vessels are often multilevel, open-air ships which require more hands working on deck and water management39.
Using LSBCI index data from 2020, we developed a network to highlight how coastal African nations are connected through maritime trade (Fig. 4). The role of this network analysis is two-fold, (1) it demonstrates an understanding of intracontinental maritime connectivity; and (2) it highlights the top three countries connected via maritime trade through an interactive html model (Supplemental File). For example, if An. stephensi is detected and established in a specific coastal African nation such as Djibouti, selecting the Djibouti node reveals the top three locations at risk of introduction from that source country (Djibouti links to Sudan, Egypt and Kenya). This can be used as an actionable prioritization list for surveillance if An. stephensi is detected in any given country and highlights major maritime hubs in Africa which could be targeted for surveillance and control. For example, since the development of this model, An. stephensi has been detected in Nigeria. Through the use of this interactive model, Ghana, Cote d’Ivoire, and Benin have been identified as countries most connected to Nigeria through maritime trade and therefore surveillance prioritization activities could consider these locations.
The network analysis reveals the significance of South African trade to the rest of the continent. Due to the distance, South Africa did not appear to be high in risk of An. stephensi introduction. However, this analysis does reveal that if An. stephensi were to enter nearby countries, it could very easily be introduced because of its high centrality. Western African countries such as Ghana, Togo, and Morocco are also heavily connected to other parts of Africa. Interestingly, Mauritius appears to be highly significant to this network of African maritime trade. Based on 2020 maritime data, Mauritius is ranked as the country with the third greatest likelihood of introduction of An. stephensi and has the second highest centrality rank value of 0.159. Considering these factors, Mauritius could serve as an important port of call connecting larger ports throughout Africa or other continents. With long standing regular larval surveillance efforts across the island for Aedes spp., this island nation is well suited to look for Anopheles larvae as part of Aedes surveillance efforts for early detection and rapid response to prevent the establishment of An. stephensi. If An. stephensi were to become established in countries with high centrality ranks, further expansion on the continent could be accelerated drastically. These ports could serve as important watchpoints and indicators of An. stephensi’s incursion into Africa. Anopheles stephensi is often found in shared habitats with Aedes spp. and a great opportunity exists to leverage Aedes arboviral surveillance efforts to initiate the search for An. stephensi, especially in countries that have high potential of introduction through maritime trade.