Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Elevation as a proxy for mosquito-borne Zika virus transmission in the Americas

  • Alexander G. Watts ,

    wattsa@smh.ca

    Affiliation Li Ka Shing Knowledge Institute, St. Michael's Hospital, Toronto, Canada

    ORCID http://orcid.org/0000-0002-2675-3460

  • Jennifer Miniota,

    Affiliation Li Ka Shing Knowledge Institute, St. Michael's Hospital, Toronto, Canada

  • Heather A. Joseph,

    Affiliation Division of Global Migration and Quarantine, National Center for Emerging and Zoonotic Infectious Diseases, Centers for Disease Control and Prevention, Atlanta, Georgia, United States of America

  • Oliver J. Brady,

    Affiliation Centre for the Mathematical Modelling of Infectious Diseases and Department of Infectious Disease Epidemiology, London School of Hygiene & Tropical Medicine, London, United Kingdom

  • Moritz U. G. Kraemer,

    Affiliations Spatial Ecology and Epidemiology Group, Department of Zoology, University of Oxford, Oxford, United Kingdom, Oxford Big Data Institute, Li Ka Shing Centre for Health Information and Discovery, University of Oxford, Oxford, United Kingdom

  • Ardath W. Grills,

    Affiliation Division of Global Migration and Quarantine, National Center for Emerging and Zoonotic Infectious Diseases, Centers for Disease Control and Prevention, Atlanta, Georgia, United States of America

  • Stephanie Morrison,

    Affiliations Division of Global Migration and Quarantine, National Center for Emerging and Zoonotic Infectious Diseases, Centers for Disease Control and Prevention, Atlanta, Georgia, United States of America, Eagle Medical Services, LLC for Division of Global Migration and Quarantine, Centers for Disease Control and Prevention, Atlanta, Georgia, United States of America

  • Douglas H. Esposito,

    Affiliation Division of Global Migration and Quarantine, National Center for Emerging and Zoonotic Infectious Diseases, Centers for Disease Control and Prevention, Atlanta, Georgia, United States of America

  • Adriano Nicolucci,

    Affiliation Li Ka Shing Knowledge Institute, St. Michael's Hospital, Toronto, Canada

  • Matthew German,

    Affiliation Li Ka Shing Knowledge Institute, St. Michael's Hospital, Toronto, Canada

  • Maria I. Creatore,

    Affiliations Li Ka Shing Knowledge Institute, St. Michael's Hospital, Toronto, Canada, Dalla Lana School of Public Health, University of Toronto, Toronto, Canada

  • Bradley Nelson,

    Affiliation Division of Global Migration and Quarantine, National Center for Emerging and Zoonotic Infectious Diseases, Centers for Disease Control and Prevention, Atlanta, Georgia, United States of America

  • Michael A. Johansson,

    Affiliation Division of Vector-Borne Diseases, Centers for Disease Control and Prevention, San Juan, Puerto Rico

  • Gary Brunette,

    Affiliation Division of Global Migration and Quarantine, National Center for Emerging and Zoonotic Infectious Diseases, Centers for Disease Control and Prevention, Atlanta, Georgia, United States of America

  • Simon I. Hay,

    Affiliations Oxford Big Data Institute, Li Ka Shing Centre for Health Information and Discovery, University of Oxford, Oxford, United Kingdom, Institute for Health Metrics and Evaluation, University of Washington, Seattle, Washington, United States of America

  • Kamran Khan ,

    Contributed equally to this work with: Kamran Khan, Marty Cetron

    Affiliations Li Ka Shing Knowledge Institute, St. Michael's Hospital, Toronto, Canada, Department of Medicine, Division of Infectious Diseases, University of Toronto, Toronto, Canada

  •  [ ... ],
  • Marty Cetron

    Contributed equally to this work with: Kamran Khan, Marty Cetron

    Affiliation Division of Global Migration and Quarantine, National Center for Emerging and Zoonotic Infectious Diseases, Centers for Disease Control and Prevention, Atlanta, Georgia, United States of America

  • [ view all ]
  • [ view less ]

Elevation as a proxy for mosquito-borne Zika virus transmission in the Americas

  • Alexander G. Watts, 
  • Jennifer Miniota, 
  • Heather A. Joseph, 
  • Oliver J. Brady, 
  • Moritz U. G. Kraemer, 
  • Ardath W. Grills, 
  • Stephanie Morrison, 
  • Douglas H. Esposito, 
  • Adriano Nicolucci, 
  • Matthew German
PLOS
x

Abstract

Introduction

When Zika virus (ZIKV) first began its spread from Brazil to other parts of the Americas, national-level travel notices were issued, carrying with them significant economic consequences to affected countries. Although regions of some affected countries were likely unsuitable for mosquito-borne transmission of ZIKV, the absence of high quality, timely surveillance data made it difficult to confidently demarcate infection risk at a sub-national level. In the absence of reliable data on ZIKV activity, a pragmatic approach was needed to identify subnational geographic areas where the risk of ZIKV infection via mosquitoes was expected to be negligible. To address this urgent need, we evaluated elevation as a proxy for mosquito-borne ZIKV transmission.

Methods

For sixteen countries with local ZIKV transmission in the Americas, we analyzed (i) modelled occurrence of the primary vector for ZIKV, Aedes aegypti, (ii) human population counts, and (iii) reported historical dengue cases, specifically across 100-meter elevation levels between 1,500m and 2,500m. Specifically, we quantified land area, population size, and the number of observed dengue cases above each elevation level to identify a threshold where the predicted risks of encountering Ae. aegypti become negligible.

Results

Above 1,600m, less than 1% of each country’s total land area was predicted to have Ae. aegypti occurrence. Above 1,900m, less than 1% of each country’s resident population lived in areas where Ae. aegypti was predicted to occur. Across all 16 countries, 1.1% of historical dengue cases were reported above 2,000m.

Discussion

These results suggest low potential for mosquito-borne ZIKV transmission above 2,000m in the Americas. Although elevation is a crude predictor of environmental suitability for ZIKV transmission, its constancy made it a pragmatic input for policy decision-making during this public health emergency.

Introduction

In February 2016, at the onset of the Zika virus (ZIKV) epidemic in the Americas, governments were required to balance protecting the health of their citizens travelling abroad against their obligation to minimize unnecessary disruption to international travel and trade as per the 2005 International Health Regulations [1]. For example, when previously unaffected countries report locally-acquired, mosquito-borne cases of ZIKV, the U.S. Centers for Disease Control and Prevention (CDC) designates the country as having local active transmission and publishes a travel notice on the CDC Traveler’s Health website. While sub-national travel notices are unusual, many ZIKV affected countries are popular U.S. travel destinations with millions of annual visitors. Since a national-level travel notice could have unnecessary economic consequences, the creation of a pragmatic method to identify areas at low risk of ZIKV transmission was essential. Hence, policymakers were pressed to identify areas that were ecologically unsuitable for mosquito-borne ZIKV transmission, to refine where travel notices should not apply.

Delineating where ZIKV risks are negligible within the Americas presents a public health policy challenge as the risk of mosquito-borne arbovirus transmission is heterogeneous [2]. Because of the high proportion of individuals without symptoms or with subclinical illness, limited access to timely laboratory diagnostics, and suboptimal surveillance infrastructure in many countries, establishment of the precise locations of ZIKV transmission is challenging [3]. In addition, the reported distribution of the primary mosquito vector, Aedes aegypti (Ae. aegypti) is limited because surveillance and reporting of mosquito presence and abundance is often inconsistent within and across nations [4]. Combinations of ecological factors have been used to predict local Ae. aegypti occurrence and thus allow more precise geographic risk estimates after the introduction of ZIKV [5]; however, since the life cycle of this mosquito depends on interacting environmental factors, many cities may not fulfill necessary temperature, precipitation, and vegetative conditions to support fundamental developmental requirements of Ae. aegypti.

Elevation is a potential proxy for Ae. aegypti range as it could be more readily understood and operationalized than time-dependent meteorological factors. Elevation is an appealing environmental proxy for Ae. aegypti range because it is correlated to a variety of fundamental dynamic ecological factors critical for mosquito development, especially temperature [6], within latitudes favorable for arbovirus transmission. Elevation itself has no known direct effect on virus transmission but could be used for policy decisions as a proxy for the critical dynamic factors that influence virus transmission.

Currently, the World Health Organization uses elevation as a factor to inform travelers on the risk of yellow fever virus (YFV) acquisition, excusing the recommended vaccination for travelers whose itineraries are limited to areas above 2,300m in some African and South American regions [7,8]. To inform time-sensitive policy decisions on sub-national travel notices at the onset of the ZIKV epidemic in the Americas, we sought to identify an elevation threshold where the probability of Ae. aegypti occurring and the associated predicted risk of ZIKV infection become negligible.

Methods

To determine the probability of Ae. aegypti occurrence as a prerequisite for ZIKV transmission at varying elevations, we performed three geospatial analyses. For each 100-m elevation interval, we determined (a) the proportion of a country’s total land area where Ae. aegypti is predicted to occur above each interval, using an established species distribution modeling approach [4], (b) the size of the population living in areas where Ae. aegypti occurrence is predicted to occur above each elevation interval, derived by high resolution population maps, and (c) the number of historic human dengue cases reported above each interval, as an indicator of the longer-term extent of ZIKV transmission, assuming it follows previous patterns of expansion to dengue, a related arbovirus [9,10]. We examined the land, population, and reported case distributions of dengue to assess the feasibility of establishing an elevation threshold where the predicted risk of ZIKV transmission becomes negligible.

Data sources

Elevation.

We used the 30-arc-second (1 x 1 km) spatial resolution global multi-resolution terrain elevation data [11], available from the U.S. Geological Survey. Previous studies report Ae. aegypti elevation maxima in the Americas at 1,600m (moderately abundant) and 2,100m (present but rare) (Lozano-Fuentes et al. 2012). However, dengue risk has also been reported to be unlikely at elevations above 1,500m [12]. Based on these observed elevation range limits of Ae. aegypti and dengue, this analysis included countries in the Americas with local ZIKV transmission as of February 25, 2016 that have any areas at elevations of 1,500m or greater. This yielded 16 countries with local ZIKV transmission for analysis (Table 1).

thumbnail
Table 1. Elevations and population counts for 16 Zika-affected countries with elevations greater than 1,500 m [20].

https://doi.org/10.1371/journal.pone.0178211.t001

Predicted occurrence of Aedes aegypti.

We quantified the amount of land area with predicted Ae. aegypti occurrence using a 5 × 5 km resolution global map of the modeled distribution of Ae. aegypti [4]. This ecological niche model predicts the global distribution of Ae. aegypti by combining spatially-explicit, temperature-dependent ranges of the vector based on fundamental limits of mosquito development; geo-located and confirmed Ae. aegypti occurrence points; as well as environmental covariates (i.e., vegetation, precipitation, and urban land cover) that further explain the mosquito species distribution [4]. To increase model output specificity, we reclassified the range of probabilities of predicted Ae. aegypti occurrence (range: 0–0.99; ‘0’ being lowest and ‘0.99’ being highest) as: ‘absence’ (i.e., all values within the range: 0–0.49 reclassified to ‘0’) or ‘presence’ (i.e., all values within the range: 0.5–0.99 reclassified to ‘1’), resulting in a binary raster of ‘0’ (absence) and ‘1’ (presence). For this study, we assumed that Ae. albopictus is a less competent vector relative to Ae. aegypti for human ZIKV transmission in nature [13] as previously shown for dengue [14] and restricted our analyses to Ae. aegypti.

Human population.

To quantify the estimated population where Ae. aegypti is predicted to occur within each elevation range, we used the LandScan (2014)TM high resolution global population distribution database [15]. LandScan uses Geographic Information Systems and remote sensing technology to model average population counts over a 24-hour period for every 1x1 km area, globally.

Historic cases of dengue.

DENV is another flavivirus transmitted by Ae. aegypti mosquitoes. We assumed that the geographic extent of ZIKV transmission would be similar to that observed for DENV, based on evidence that temperature-dependent constraints on mosquito survival and associated capability to spread arboviruses to humans is similar across these and other related flaviviruses [16]. In addition, recent studies show that viral dynamics of DENV and ZIKV in humans are similar [17]. To estimate the elevation range where dengue cases are negligible, we used a global geographic database of dengue cases between 1960–2012 (N = 8,309, [18]). This is the most comprehensive database of confirmed dengue cases with enough detailed information to carry out modelling at a sub-national scale. Cases are represented as occurrence data, linked to point or polygon locations, derived from peer-reviewed literature, case reports, or informal online sources.

Statistical analysis

We reclassified the elevation spatial data to represent land area as vertical elevation intervals. The vertical elevation intervals were divided into three classes: (1) elevations between 0–1,000m; (2) elevations between 1,000m and 2,500m, subsequently divided into fifteen 100m ranges; and (3) elevations greater than 2,500m, to a maximum of 8,800m. This resulted in 17 elevation ranges used to estimate land area with predicted Ae. aegypti occurrence, human population counts in those areas, and historic human dengue cases. We chose to further divide the second elevation class by fifteen 100m intervals between 1,000m and 2,500m to ensure that we captured any potential elevation level that might fall between the estimated range of Ae. aegypti that has been previously observed in the Americas (i.e., between 1,700m and 2,100m [19]).

We quantified the land area where Ae. aegypti is predicted to occur within each elevation range per country. First, we multiplied the reclassified binary raster of predicted Ae. aegypti occurrence (5 x 5 km) [4] with the reclassified elevation raster to standardize the spatial resolution between the Ae. aegypti and elevation models for each 100m increment elevation interval. To evaluate an elevation interval above which Ae. aegypti is expected to occur or not occur with very rare frequency, we divided the aggregated sum of land area classified as Ae. aegypti-‘present’ above each elevation interval by the total land area of each country with local ZIKV transmission. These estimates resulted in a value for the remaining land area (%) where there is predicted Ae. aegypti occurrence above each elevation level, for each analyzed country.

To evaluate the population at risk of encountering Ae. aegypti for the 16 analyzed countries with local ZIKV transmission, we selected all LandScan pixels that intersected the binary Ae. aegypti occurrence raster within each elevation range using zonal statistics (i.e., measures of descriptive statistics for a spatial entity within a given geographic area). We then summed the estimated population within those pixels. We compared the population at risk of encountering Ae. aegypti above all elevation intervals for each analyzed country, assuming that populations residing in an area predicted to have Ae. aegypti occurrence puts them at risk of encountering Ae. aegypti.

We used historic dengue occurrence points to validate the likelihood of ZIKV transmission at varying elevations. We first selected all dengue cases (n = 2,950) reported from the 16 analyzed countries. We removed 248 occurrence points (8.4% of total cases) where the location of occurrence could not be confirmed at a city or point level of geographic resolution (e.g. those that were reported only at state or provincial level) due to high variability in elevation (and therefore uncertainty of the specific location of DENV infection, within those states or provinces). We then counted the remaining geo-positioned dengue cases (n = 2,702) that were located within each elevation interval. Finally, we calculated the proportion of all reported human dengue cases reported above each elevation interval within each of the 16 analyzed countries.

Using this elevation-specific modeled Ae. aegypti distribution map, population counts, and geo-located dengue cases for the 16 countries with local ZIKV transmission, we defined three criteria for choosing an elevation threshold where the risk of ZIKV exposure by Ae. aegypti mosquitoes becomes negligible: (a) where the average proportion of land area above a given elevation with predicted Ae. aegypti occurrence is approximately 1% or less of total land area of the selected country or territory; (b) the proportion of the human population living above a given elevation and where Ae. aegypti is predicted to occur is approximately 1% or less of the total population of the selected country or territory; and (c) the proportion of dengue case points reported above a given elevation is approximately 1% or less of the total number of reported dengue case points for each selected country or territory. We chose 1% as the criterion for rare occurrence of Ae. aegypti and associated disease risks to account for possible error in the Ae. aegypti distribution model and reporting biases in the dengue case points. To be conservative, we strictly chose the maximum elevation threshold that met any one of these three criteria.

Results

During the initial stages of the ZIKV epidemic in the Americas (i.e. as of February 26, 2016), 16 of 36 countries with reported local ZIKV transmission were deemed to have physical geographic features at or above 1,000m. Table 1 displays the mean and maximum elevations and total population in each of the 16 countries, and the population and elevation of the most populous cities in each of these countries [20].

At higher elevations, less land area was predicted to have Ae. aegypti (Fig 1). On average across the 16 countries, 1% or less land area above 1,600m was predicted to have Ae. aegypti (range: 0.02% in Nicaragua to 3.99% in Mexico). The three countries with the greatest proportion of land area above 1,600 m where Ae. aegypti occurrence was also predicted, were Mexico (3.99%), Guatemala (2.51%), and Ecuador (1.53%).

thumbnail
Fig 1. Proportion of a country’s total land area above each elevation threshold, where Ae. aegypti is predicted to occur.

https://doi.org/10.1371/journal.pone.0178211.g001

At higher elevations, the predicted risk of humans encountering Ae. aegypti also decreased. On average across the 16 countries, 1% or less of the total human population living above 1,900m remained at risk of encountering Ae. aegypti (Fig 2); these values ranged from 0.00% (Brazil, Dominican Republic, Guyana, Jamaica, Nicaragua, and Panama) to 2.52% (Bolivia). Countries with the greatest proportion of population above 1,900m and predicted to have Ae. aegypti occurrence were Bolivia (2.52%); Guatemala (1.84%); and Colombia (1.11%).

thumbnail
Fig 2. Proportion of a country’s total population living above each elevation threshold, where Ae. aegypti is predicted to occur.

https://doi.org/10.1371/journal.pone.0178211.g002

On average across the 16 countries, 1% or less of all historically reported dengue cases were observed above 2,000m (Table 2); six of the 16 analyzed countries reported some dengue cases above 2,000m. Of the reported dengue cases used for this analysis [18], Colombia had the highest number located above 2,000m (11 cases; 9.6% of all cases in Colombia) followed by Bolivia (eight cases; 9.4% of all cases in Bolivia) and Mexico (six cases; 2.6% of all cases in Mexico).

thumbnail
Table 2. Geo-located human dengue virus (DENV) cases reported by Messina et al. 2014 in the 16 Zika-affected countries for all elevation intervals (m).

https://doi.org/10.1371/journal.pone.0178211.t002

The maximum elevation that met any one of the three threshold criteria was 2,000m; therefore, we calculated Ae. aegypti coverage and corresponding populations above 2,000m. On average across the 16 countries, above 2,000m, Ae. aegypti was predicted to occur in less than 0.25% of the total land area. Countries with the greatest proportion of land area above 2,000m predicted to have Ae. aegypti occurrence were Guatemala (0.90%); Mexico (0.89%); Ecuador (0.62%); and Bolivia (0.60%). Brazil, Guyana, Nicaragua, and Panama were predicted to have no Ae. aegypti occurrence above 2,000m. On average across the 16 countries, less than 0.28% of total population living above 2,000m was estimated to be at risk of encountering Ae. aegypti. In terms of human population distribution, countries with the largest populations living above 2,000m and in areas where Ae. aegypti was predicted to occur were Colombia (441,836; 1.0% of total Colombian population); Mexico (416,113; 0.4% of total Mexican population, Fig 3); Bolivia (240,759; 2.3% of total Bolivian population); and Guatemala (230,840; 1.6% of total Guatemalan population) (Tables 1 and 3). Conversely, Mexico City, Mexico (8.8 million); Bogota, Colombia (7.7 million); Puebla, Mexico (1.4 million); Quito, Ecuador (1.4 million, Fig 4); and Cochabamba, Bolivia (0.9 million), all above 2,000m, were not among cities predicted to have Ae. aegypti (Figs 3 and 4).

thumbnail
Fig 3. Mexico’s land area above 2,000m where Ae. aegypti is predicted to occur.

https://doi.org/10.1371/journal.pone.0178211.g003

thumbnail
Fig 4. Ecuador’s land area above 2,000m where Ae. aegypti is predicted to occur.

https://doi.org/10.1371/journal.pone.0178211.g004

thumbnail
Table 3. Proportions of land area and populations with predicted occurrence of Ae. aegypti above 2,000 m for 16 Zika-affected countries in descending order of human population above 2,000m where Ae. aegypti is predicted to occur.

https://doi.org/10.1371/journal.pone.0178211.t003

Discussion

Our analysis demonstrated that elevation is a crude yet practical proxy for the presence of Ae. aegypti, and consequently the risk of acquiring Zika virus infection, in the Americas. We identified that, above 1,600m and 1,900m, the predicted occurrence of Ae. aegypti and associated human risks of encountering Ae. aegypti were greatly diminished, respectively. We further demonstrated that human cases of DENV were rarely reported above 2,000m in this region of the world. Hence, using the maximum elevation that met all three of our threshold criteria, our analysis suggests that above 2,000m in the Americas environmental conditions are poorly suited for the transmission of ZIKV via Ae. aegypti.

While a single static factor such as elevation does not capture all dynamic processes that influence the spatial and temporal extent of mosquito-borne disease risks [21], we believe it is a pragmatic proxy for Ae. aegypti range because it is correlated to a variety of fundamental dynamic ecological factors critical for mosquito development and transmission of disease. For example, our results rely heavily on model outputs of the global predicted occurrence of Ae. aegypti [4] which uses a temperature-driven suitability filter defining the fundamental limits of Ae. aegypti, in addition to other model predictors of Ae. aegypti species range [5,22]. Vectorial capacity, defined as the average rate at which potentially infective mosquito bites arise following the introduction of a single infectious host, is highly influenced by extrinsic environmental factors [23,24]. Temperature and humidity highly influence the probability of daily survival and gonotrophic period for successful oviposition of eggs [6,25]. Because Ae. aegypti suitability is heavily temperature-dependent [4], it is likely that above 2,000m the temperature-dependent capacity of Ae. aegypti mosquitoes to survive long enough to transmit ZIKV to humans is limited.

Topography, notably elevation, has previously been used to identify vertical thresholds in mosquito-borne pathogen transmission [19,26,27]. Thermal constraints on adult mosquitoes have precluded the successful development and occurrence of Ae. aegypti in some high-elevation regions in south-east Australia [28]. Estimates of elevation maxima of Ae. aegypti in Mexico report that Ae. aegypti was commonly observed up to 1,700m and present but rare from 1,700m to 2,130m [19]. In Peru, researchers reported small dengue outbreaks that occurred at elevations up to approximately 1,500m [29]. However, in Colombia, Aedes spp. mosquitoes have been reported to survive a life-cycle indoors at elevations as great as 2,200m [30]. Together, these findings support the assertion that above 2,000m sustained Ae. aegypti populations are only partially supported by human dwellings above this elevation level. With increased surveillance capacity over the tropical Americas, especially in well-connected cities at high elevations such as La Paz (Bolivia), Quito (Ecuador), and Bogota (Colombia), our model would be improved by re-evaluating the elevation thresholds above which ongoing observed occurrences of both Ae. aegypti and confirmed cases of DENV and ZIKV are reported. Additionally, modeling the predicted occurrence of ZIKV, parameterized by improved measures of the temperature-dependent constraints on ZIKV transmission to humans, could help identify rare locations above 2,000m at risk of ZIKV transmission. While the purpose of this analysis was to provide advice more generally regarding areas more or less likely to be at risk of ZIKV transmission, travelers should always consider local conditions prior to travel as exceptions do occur.

Our study has several limitations. In the Ae. aegypti model [4], all environmental covariates used were annual summaries of seasonal conditions. As a result, our analysis does not account for seasonality variability in suitability for ZIKV transmission. Many areas within the 1,600–1,900m elevation range are likely to experience only limited windows of time during which Ae. aegypti might persist. Thus, especially in the winter time, 2,000m might be an overly conservative elevation threshold beyond which ZIKV transmission might occur [31].

Furthermore, we assumed that DENV occurrences reflect potential ZIKV occurrences at similar elevations. This assumption is limited given that specific temperature-dependent constraints on DENV transmission to humans closely resemble–but are not perfectly aligned–to those conditions essential for ZIKV transmission to humans [14]. However, regional geographic distributions and recent spread of ZIKV has followed that of DENV [32,33], and such strong similarities in the transmission characteristics between these arboviruses support our assumption without confirmed ZIKV cases in the study region at the time of investigation.

Our analysis of dengue cases in the 16 countries was performed on a historic sample dataset of geo-positioned reported DENV occurrences. This dataset is limited to observations sampled in 1960–2012 and therefore does not capture cases reported after 2012 (including cases related to high incidence of DENV in the Americas during 2015 [34]). Given more updated DENV occurrences, our elevation threshold could potentially increase depending on the location of the confirmed cases. In the available historical dataset, many cases of dengue were reported in our study region but it is likely that some cases were left unreported in particular locations due to lack of reporting or laboratory infrastructure [18]. Additionally, reporting biases in the dengue dataset limit the true spatial location of encounter by a human host with the DENV-infected mosquito vector. Finally, recent evidence suggests that ZIKV geographic range might be more restricted than that of DENV [35].

This study focused on potential ZIKV transmission by the primary vector Ae. aegypti, but did not consider transmission by other Aedes mosquito species, such as Ae. albopictus. Historical observations of ZIKV in Ae. albopictus mosquitoes [36] and more recent PCR-detected ZIKV infection of Ae. albopictus mosquitoes captured in the environment in Mexico [37], suggests that this species may be a competent vector for ZIKV transmission. However, Ae. albopictus is often considered a less effective vector in ZIKV transmission because Ae. aegypti feeds on human hosts more frequently than Ae. albopictus mosquitoes [38,39] and is generally more susceptible to ZIKV than Ae. albopictus [9,25]. In the case of dengue transmission, Ae. albopictus has been shown to have a limited role in population-level transmission even in highly endemic settings [14]. Therefore, while Ae. albopictus may be tolerant to cooler temperatures at higher elevations relative to Ae. aegypti and may transmit ZIKV at those elevations [6,25], it is unlikely that the potential impact of Ae. albopictus on ZIKV transmission was more critical than Ae. aegypti, especially at higher elevations. The extent of the 2015–2016 Zika outbreak appears to support these decisions as no transmission was reported in the USA in areas where Ae. albopictus exists but did occur where Ae. aegypti is known to occur.

Conclusions

During the initial stages of the 2016 ZIKV epidemic in the Americas, difficult time-sensitive decisions were needed to balance the health risks of travelers to countries affected by ZIKV against the economic consequences of travel notices to the entirety of these countries. On March 11, 2016, these findings were applied to U.S. ZIKV-related travel notices informing travelers that the risk of infection in areas above 2,000m was negligible [40]. Since March 11, 2016, to our knowledge, no locally-acquired, mosquito-borne infections have occurred above 2,000m in any country or U.S. territory. However, prospective validation with data from human disease and vector surveillance is needed to determine if this elevation threshold continues to reflect a low risk of exposure to ZIKV. While dynamic meteorological data might have offered additional spatio-temporal precision in assessing the occurrence of Ae. aegypti, and consequently the sub-national risk of acquiring ZIKV infection in the Americas, we found elevation to be a pragmatic proxy to inform policy and may be readily understood by travelers.

Acknowledgments

Disclaimer: The conclusions, findings, and opinions expressed by authors contributing to this journal do not necessarily reflect the official position of the U.S. Centers for Disease Control and Prevention, or the authors' affiliated institutions. Use of trade names is for identification only and does not imply endorsement by the Public Health Service or by the U.S. Department of Health and Human Services.

We thank Carmen Huber, Mark Ong, Anne Christian, Sonya Karamchandani at Li Ka Shing Knowledge Institute, St. Michael's Hospital, Toronto, Canada, and Ryan Lash (Centers for Disease Control and Prevention, Atlanta, Georgia, USA) for spatial analyses and GIS support. We also thank Maggie Chan and Joanna Vass at the Li Ka Shing Knowledge Institute, St. Michael's Hospital, Toronto, Canada for support in figure design.

Author Contributions

  1. Conceptualization: MC KK AG AW JM HJ SM BN MG MIC.
  2. Data curation: OB MK SH.
  3. Formal analysis: AW MG AN MIC JM KK.
  4. Funding acquisition: MC KK SH.
  5. Investigation: AW JM HJ AG BN SM KK MC.
  6. Methodology: MG AN JM KK AW HJ AG MJ.
  7. Project administration: MC KK JM HJ.
  8. Software: AN OB MK SH.
  9. Supervision: JM HJ KK AG.
  10. Visualization: AW KK JM HJ.
  11. Writing – original draft: AW.
  12. Writing – review & editing: AW JM HJ AG SM BN GB DE.

References

  1. 1. World Health Organization. International Health Regulations. WHO. 2005; 82.
  2. 2. Weiss RA, McMichael AJ. Social and environmental risk factors in the emergence of infectious diseases. Nat Med. 2004;10: S70–S76. pmid:15577934
  3. 3. Woolhouse M. How to make predictions about future infectious disease risks. Philos Trans R Soc Lond B Biol Sci. 2011;366: 2045–54. pmid:21624924
  4. 4. Kraemer MUG, Sinka ME, Duda KA, Mylne AQN, Shearer FM, Barker CM, et al. The global distribution of the arbovirus vectors Aedes aegypti and Ae. albopictus. Elife. 2015;4: e08347. pmid:26126267
  5. 5. Kraemer MUG, Hay SI, Pigott DM, Smith DL, Wint GRW, Golding N. Progress and challenges in infectious disease cartography. Trends Parasitol. 2016;32: 19–29. pmid:26604163
  6. 6. Brady OJ, Golding N, Pigott DM, Kraemer MUG, Messina JP, Reiner RC, et al. Global temperature constraints on Aedes aegypti and Ae. albopictus persistence and competence for dengue virus transmission. Parasit Vectors. 2014;7: 338. pmid:25052008
  7. 7. Shlim DR, Freedman DO, Keystone JS, Tan KR, Kozarsky PE, Borwein ST, et al. CDC Health Information for International Travel 2010. In: Brunette Gary W., editor. CDC Health Information for International Travel 2010. New York: Oxford University Press; 2009. pp. 242–269.
  8. 8. Jentes ES, Poumerol G, Gershman MD, Hill DR, Lemarchand J, Lewis RF, et al. The revised global yellow fever risk map and recommendations for vaccination, 2010: Consensus of the Informal WHO Working Group on Geographic Risk for Yellow Fever. Lancet Infect Dis. 2011;11: 622–632. pmid:21798462
  9. 9. Chouin-Carneiro T, Vega-Rua A, Vazeille M, Yebakima A, Girod R, Goindin D, et al. Differential susceptibilities of Aedes aegypti and Aedes albopictus from the Americas to zika virus. PLoS Negl Trop Dis. 2016;10: e0045453.
  10. 10. Anderson JR, Rico-Hesse R. Aedes aegypti vectorial capacity is determined by the infecting genotype of dengue virus. Am J Trop Med Hyg. 2006;75: 886–892. pmid:17123982
  11. 11. Danielson J, Gesch D. Global Multi-resolution Terrain Elevation Data 2010(GMTED2010). US Geol Surv Open-File Rep 2011–1073. 2011;2010: 26.
  12. 12. Hagenlocher M, Delmelle E, Casas I, Kienberger S. Assessing socioeconomic vulnerability to dengue fever in Cali, Colombia: statistical vs expert-based modeling. Int J Health Geogr. 2013;12: 36. pmid:23945265
  13. 13. Richards SL, Ponnusamy L, Unnasch TR, Hassan HK, Apperson CS. Host-feeding patterns of Aedes albopictus (Diptera: Culicidae) in relation to availability of human and domestic animals in suburban landscapes of central North Carolina. J Med Entomol. 2006;43: 543–551. pmid:16739414
  14. 14. Lambrechts L, Scott TW, Gubler DJ, Rousseaux C, Khun H. Consequences of the expanding global distribution of Aedes albopictus for dengue virus transmission. Halstead SB, editor. PLoS Negl Trop Dis. W. H. Freeman and Co; 2010;4: e646. pmid:20520794
  15. 15. ORNL. LandScan 2013. In: Oak Ridge National Lab [Internet]. 2014. Available from: http://web.ornl.gov/sci/landscan/
  16. 16. Lambrechts L, Paaijmans KP, Fansiri T, Carrington LB, Kramer LD, Thomas MB, et al. Impact of daily temperature fluctuations on dengue virus transmission by Aedes aegypti. Proc Natl Acad Sci U S A. 2011;108: 7460–7465. pmid:21502510
  17. 17. Paz-Bailey G, Rosenberg ES, Doyle K, Munoz-Jordan J, Santiago GA, Klein L, et al. Persistence of zika virus in body fluids—Preliminary Report. N Engl J Med. 2017; NEJMoa1613108.
  18. 18. Messina J, Brady O, Pigott D, Brownstein J, Hoen A, Hay S. A global compendium of human dengue virus occurrence. Sci Data. 2014;1: 140004. pmid:25977762
  19. 19. Lozano-Fuentes S, Hayden MH, Welsh-Rodriguez C, Ochoa-Martinez C, Tapia-Santos B, Kobylinski KC, et al. The dengue virus mosquito vector Aedes aegypti at high elevation in Mexico. Am J Trop Med Hyg. 2012;87: 902–909. pmid:22987656
  20. 20. US Census Bureau. World Population: Historical Estimates of World Population [Internet]. US Census Bureau. 2013. Available: http://www.census.gov/population/international/data/worldpop/table_history.php
  21. 21. Peterson AT, Osorio J, Qiao H, Escobar LE. Zika virus, elevation, and transmission risk. PLoS Curr. 2016;8: 9–15.
  22. 22. Fuller DO, Troyo A, Beier JC. El Niño Southern Oscillation and vegetation dynamics as predictors of dengue fever cases in Costa Rica. Environ Res Lett. 2009;4: 140111–140118. pmid:19763186
  23. 23. Thu HM, Aye KM, Thein S. The effect of temperature and humidity on dengue virus propagation in Aedes aegypti mosquitos. Southeast Asian J Trop Med Public Health. 1998;29: 280–284. Available: http://www.ncbi.nlm.nih.gov/pubmed/9886113 pmid:9886113
  24. 24. Stone CM, Jackson BT, Foster WA. Effects of plant-community composition on the vectorial capacity and fitness of the malaria mosquito Anopheles gambiae. Am J Trop Med Hyg. 2012;87: 727–736. pmid:22927493
  25. 25. Brady OJ, Johansson M, Guerra C, Bhatt S, Golding N, Pigott DM, et al. Modelling adult Aedes aegypti and Aedes albopictus survival at different temperatures in laboratory and field settings. Parasit Vectors. 2013;6: 351. pmid:24330720
  26. 26. Althouse BM, Hanley KA, Diallo M, Sall AA, Ba Y, Faye O, et al. Impact of climate and mosquito vector abundance on sylvatic arbovirus circulation dynamics in senegal. Am J Trop Med Hyg. American Society of Tropical Medicine and Hygiene; 2015;92: 88–97. pmid:25404071
  27. 27. Chowell G, Cazelles B, Broutin H, Munayco C V. The influence of geographic and climate factors on the timing of dengue epidemics in Perú, 1994–2008. BMC Infect Dis. 2011;11: 164. pmid:21651779
  28. 28. Kearney M, Porter WP, Williams C, Ritchie S, Hoffmann AA. Integrating biophysical models and evolutionary theory to predict climatic impacts on species’ ranges: The dengue mosquito Aedes aegypti in Australia. Funct Ecol. 2009;23: 528–538.
  29. 29. Chowell G, Torre C a, Munayco-Escate C, Suárez-Ognio L, López-Cruz R, Hyman JM, et al. Spatial and temporal dynamics of dengue fever in Peru: 1994–2006. Epidemiol Infect. 2008;136: 1667–1677. pmid:18394264
  30. 30. Suárez MF, Nelson MJ. Registro de altitud del aedes aegypti en Colombia. Biomédica (Bogotá). 1981;1: 225.
  31. 31. Bogoch II, Brady OJ, Kraemer MUG, German M, Creatore MI, Kulkarni MA, et al. Anticipating the international spread of Zika virus from Brazil. Lancet. 2016;387: 335–336. pmid:26777915
  32. 32. Musso D, Cao-Lormeau VM, Gubler DJ. Zika virus: following the path of dengue and chikungunya? Lancet. 2015;386: 243–244. pmid:26194519
  33. 33. Fong IW. Chikungunya virus and zika virus expansion: an imitation of dengue virus. Emerging Zoonoses. Cham: Springer International Publishing; 2017. pp. 101–130.
  34. 34. WHO: Dengue and severe dengue. WHO. World Health Organization; 2017; Available: http://www.who.int/mediacentre/factsheets/fs117/en/
  35. 35. Carlson C, Dougherty E, Getz W. An ecological assessment of the pandemic threat of Zika virus. PLoS Negl Trop Dis. 2016;10:e0004968. pmid:27564232
  36. 36. Grard G, Caron M, Mombo IM, Nkoghe D, Mboui Ondo S, Jiolle D, et al. Zika virus in Gabon (Central Africa)‗2007: A new threat from Aedes albopictus? PLoS Negl Trop Dis. 2014;8.
  37. 37. Pan American Health Organization, World Health Organization. Zika—epidemiological update [Internet]. 2016. Available: http://www2.paho.org/hq/index.php?option=com_docman&task=doc_view&Itemid=270&gid=34243&lang=en
  38. 38. Delatte H, Desvars A, Bouétard A, Bord S, Gimonneau G, Vourc’h G, et al. Blood-feeding behavior of Aedes albopictus, a vector of Chikungunya on La Réunion. Vector Borne Zoonotic Dis. 2010;10: 249–258. pmid:19589060
  39. 39. Thavara U, Tawatsin A, Chansang C, Kong-ngamsuk W, Paosriwong S, Boon-Long J, et al. Larval occurrence, oviposition behavior and biting activity of potential mosquito vectors of dengue on Samui Island, Thailand. J Vector Ecol. 2001;26: 172–180. pmid:11813654
  40. 40. Cetron M. Revision to CDC’s zika travel notices: minimal likelihood for mosquito-borne zika virus transmission at elevations above 2,000 meters. MMWR Morb Mortal Wkly Rep. 2016;65: 267–8. pmid:26985965