An outbreak of Rocky Mountain spotted fever (RMSF) transmitted by the brown dog tick (Rhipicephalus sanguineus sensu lato) has emerged as a major human and animal health concern in Mexicali, Mexico. Due to high rates of brown dog tick infestation, susceptibility, and association with humans, dogs serve as sentinels and have a key role in the ecology of RMSF. A cross-sectional household questionnaire study was conducted in six rural and urban locations to characterize dog ecology and demography in RMSF high-and low-risk areas of Mexicali. In addition, we tracked movement patterns of 16 dogs using a GPS data logger. Of 253 households, 73% owned dogs, and dog ownership tended to be higher in high-risk areas, with a mean dog:human ratio of 0.43, compared with 0.3 in low-risk areas. Dogs in high-risk areas had higher fecundity and roamed more, but the dog density and numbers of free-roaming dogs were comparable. There was a higher proportion of younger dogs and lower proportion of older dogs in high-risk areas. The high proportion of immunologically naïve puppies in high risk areas could result in a lack of herd immunity leading to a more vulnerable dog and human population. The marked increase of space use of free-roaming dogs in high-risk areas suggests that unrestrained dogs could play an important role in spreading ticks and pathogens. As means to limit RMSF risk, practical changes could include increased efforts for spay-neuter and policies encouraging dog restraint to limit canine roaming and spread of ticks across communities; due to dog density is less impactful such policies may be more useful than restrictions on the number of owned dogs.
Citation: López-Pérez AM, Orozco L, Zazueta OE, Fierro M, Gomez P, Foley J (2020) An exploratory analysis of demography and movement patterns of dogs: New insights in the ecology of endemic Rocky Mountain-Spotted Fever in Mexicali, Mexico. PLoS ONE 15(5):
Editor: Roman R. Ganta, Kansas State University, UNITED STATES
Received: February 5, 2020; Accepted: May 7, 2020; Published: May 21, 2020
Copyright: © 2020 López-Pérez et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data are within the manuscript.
Funding: JEF and ALP acknowledge funding support from the Pacific Southwest Regional Center of Excellence for Vector-Borne Diseases funded by the U.S. Centers for Disease Control and Prevention (Cooperative Agreement 1U01CK000516).
Competing interests: The authors have declared that no competing interests exist.
Domestic dogs (Canis lupus familiaris) have featured in human civilization for tens of thousands of years, providing protection, assistance hunting, companionship, and other services [1, 2]. However, dogs share pathogens and parasites with humans, in some cases bridging disease from other animals including wildlife to people, serving as reservoirs for shared disease agents, and acting as sentinels for human risk . Dogs with greater roles in human disease ecology include those that interact with wildlife or roam free, those in populations characterized by rapid turnover (due to high mortality or fecundity rates), and those subject to low standards of veterinary care including vaccination. Examples of diseases associated with pet dogs include rabies, which is endemic in some areas where dogs are not restrained or vaccinated , and the tapeworm Echinococcus for which dogs and other canids can be important sources of environmental contamination .
Rocky Mountain spotted fever (RMSF), a potentially fatal disease of dogs and people caused by the bacterium Rickettsia rickettsii, is the most important rickettsiosis in North America [6, 7]. Tick vectors of RMSF include the American dog tick (Dermacentor variabilis), the wood tick (D. andersoni), the brown dog tick (Rhipicephalus sanguineus sensu lato), as well as several Amblyomma tick species (e.g. A. cajennense, A. imitator, and A. parvum) [8–12]. While Dermacentor spp. ticks are essentially sylvatic, preferring wild small mammal hosts in immature stages, all feeding stages of the brown dog tick prefer dogs, such that this tick is unusual in being fully peri-domestic. The abundance of the brown dog tick is determined by the number of dogs, especially stray dogs, in a community [8, 13, 14], with ticks sometimes spilling over to feed on humans. Rickettsial infections in dogs influence prevalence in ticks and serve to amplify the presence of the pathogen [8, 15].
Epidemics of RMSF have been associated with high numbers of stray dogs and uncontrolled tick infestations in eastern Arizona in the United States and in the Mexican states of Sonora and Baja California [8, 13]. At least 250 cases with 19 deaths occurred among American Indians in Arizona from 2003–2012, prompting an aggressive and successful response by public health officials incorporating long-acting acaricidal collars on dogs, treatment of the environment, and spay-neuter plus tethering of dogs . The RMSF epidemic in Baja California is much larger than those in Arizona. Since 2008, an outbreak of RMSF is ongoing in Mexicali, a city of 700,000 people immediately south of the border with the US, affecting at least 1000 people and countless dogs, with a high human mortality rate of approximately 40% . In Mexico, the vector is a distinct tropical lineage of brown dog tick, in contrast with epidemics in Arizona associated with the temperate lineage ; increased warming and drying associated with climate change may permit the tropical lineage ticks to expand northward, with evidence that canine exposure is increased directly north of the US/Mexico border .
Epidemiological determinants of RMSF in Mexico are not well-understood. Ticks can acquire the infection from infected dogs as well as transovarially through passage of the bacterium through tick eggs [19, 20]. However, surveillance may reveal very low prevalence of antibodies in dogs and low proportion of Rickettsia-positive ticks , possibly due in part to fine spatial scale endemic foci as well as local disease extinction and introduction of disease to new areas. Across Mexicali, there appears to be a higher risk for human cases and seropositive dogs in some neighborhoods on the periphery of the city and in agricultural small towns (ejidos) in the valley that extends to the southeast of the city [21–23]. Even though local elevations in tick numbers increase risk, there have been cases of RMSF and documented brown dog tick infestations even in homes without dogs . Although RMSF may be fatal in dogs, animals that survive infection may develop immunity for some period of time, suggesting that abundant puppies could reduce herd immunity and enhance the persistence of the pathogen in a community. Human decisions about dog ownership including allowing animals to roam, not treating the dogs with acaricides, and keeping dogs outside could all impact tick infestation risk as well as risk of RMSF both in the dog and in humans in the household.
Our goals in this study were to document dog population size, dog:human ratios, and demographic characteristics of canine populations in neighborhoods characterized as high and low-risk for RMSF, and to evaluate canine movement patterns in these differing areas.
Dog Materials and methods
Work was approved by the University of California Davis Institutional Animal Care and Use Committee (protocol # 20483) and the Hospital General de Mexicali Research Committee protocol # 02-01-HGMXL/ISESALUD/CDC/UCDAVIS-2019-08-07-250.
The study was conducted during August 2019 (when ticks have increased abundance and activity ) at six locations, including three rural villages and three urban areas in and near Mexicali in northwestern Mexico (Fig 1). Mexicali is located within the lower Colorado River Delta in the Sonoran Desert, and comprises a mosaic of landscapes dominated by urban and rural settlements and agriculture, with some areas of riparian and shrub-dominated vegetation. The climate is arid with a mean annual precipitation of 55.4 mm and highly variable temperatures from 6°C—42°C with a mean annual temperature of 22.4 °C.
Fig 1. Sampling sites of rural and urban areas in Mexicali Valley, Mexico.
Capital letters refer to locations (Urban neighborhoods: GM: Gabriela Mistral, HP Hacienda Los Portales, VC: Venustiano Carranza; rural areas: ED: Ejido Durango, EO: Ejido Oaxaca, AG: Algodones). Reprinted from the OpenStreetMap vector basemap hosted by Environmental Systems Research Institute (Esri) and provided under a CC BY 4.0 license, (Map data © OpenStreetMap contributors, Map layer by Esri 2019).
The locations chosen for study included two urban and two rural sites that were categorized as high-risk for RMSF (Gabriela Mistral, 32°37’N, 115°34’W, and Hacienda Portales, 32°35’N, 115°29’W in Mexicali City and Ejido Durango, 32°14’N, 115°15’W and Algodones, 32°42’N, 114°44’W in the agricultural valley), and one low-risk rural and urban control (Venustiano Carranza, 32°36’N, 115°24’W in Mexicali City and Ejido Oaxaca, 32°21’N, 115°10’W in the valley). RMSF risk was classified based on a previous study , defining a high-risk area as the one that meet three criteria 1) at least one confirmed R. rickettsii-positive tick; 2) a prevalence of RMSF antibodies in dogs > 30% (cut-off titer >1:64); and 3) a confirmed human case of RMSF within 12 months preceding the study.
A cross-sectional household questionnaire was conducted. We randomly selected three blocks from each rural town and urban neighborhood. At each block, questionnaires were deployed by choosing a random house and asking if the inhabitants consented to participate. If consent was not provided or the household did not own a dog (s), the team continued around the block and the surrounding areas until obtaining consent of at least 10 dog-owning households. Questionnaires were delivered orally in Spanish by students from the Medical School of Baja California Autonomous University and personnel of the Baja California Secretary of Health. Questions focused on number of dogs, demography of dogs, female dog reproductive history, and household dog ownership (S1 Table). At each sampled household, the person at home and best able to answer questions related to dogs and their care was asked for a verbal consent to participate in the survey. Institutional review board (IRB) approval was not required as all questions specifically related to care and husbandry of the dogs and were covered under the veterinarian-client relationship.
Households were classified as dog-owning (DOHH) and non-dog owning (NDOHH). For DOHH, we recorded the number of dogs in the household, and we asked the owners questions relating to reasons for owning the dog(s) (guardian, shepherd, pet, hunting), origin of dogs (acquired from neighbor or family, bought, found or adopted, or born at home), number of sterilized animals, the presence of free-roaming dogs of unknown origin and if each animal was allowed to roam freely. We also asked how many dogs were female and how many litters had been produced by female dogs. The area of each neighborhood and rural village was determined by drawing a perimeter of the polygon using QGIS 3.4 (QGIS Development Team, 2009).
GPS-tracking data collection
Sixteen free-roaming male dogs were collared and monitored from August 10–22, 2019: these included five dogs (three were 1-yr old, one was 4 yr-old, and one was 5-yr old) from a high-risk rural village (Ejido Durango), five (ages 1-yr for two dogs and one dog each aged 2, 3, and 4-yr old) from low-risk rural Ejido Oaxaca, three (two 2-yr old and one 6-yr old) from high-risk urban Colonia Hacienda Portales and three (ages 1, 2, and 5-yr old) from low-risk urban Colonia Venustiano Carranza. None of these dogs was neutered. All collaring and tracking was done after obtaining written informed consent from the owners. Collars were lightweight nylon with an attached motion-detecting GPS-data logger (Igot-U GT600, Mobile Action Technology, Taipei, Taiwan) configurated to record a location every 10 min. Logger data were downloaded into the i-gotU Sport Analyzer software (Mobile Action Technology). Each track was visually inspected for anomalies and gaps in data collection, and GPS coordinates for each recorded location were extracted into a spreadsheet. Locations were visualized and analyzed with QGIS. We estimated the area occupied (AO) by each dog by calculating the minimum convex polygon (MCP) with the Minimum Bounding Geometry tool. We determined the number of houses that each dog visited (NHV) as the number of houses within each occupied area. Maximum distance moved (MDM) by each dog was calculated as the maximum distance among location points using the Measure tool.
Summary statistics were calculated including mean number of dogs per DOHH for each neighborhood. We estimated the dog population size and dog density for each site using the data derived from the questionnaires (proportion of households that own dogs and mean numbers of dogs per household). The dog population size was calculated by multiplying the number of inhabited households by the proportion of dog owning household and then by the mean number of dogs per household that owned at least one dog in each neighborhood and rural village. The dog population density was calculated by dividing the estimated dog population by the area of each site. The dog:human ratio was estimated by dividing the total human population by the total number of dogs calculated for each location. Human population size and the number of inhabited households were taken from the human census of 2016 performed by the Instituto Nacional de Estadística y Geografia of Mexico . Female reproduction information was used to calculate mean litter size and female fecundity. Fecundity was determined as the number of female offspring per female per year m(x), i.e. proportion of breeding females (>12months) and the mean number of pups/female in the past 12 months following . This fecundity estimation assumes a 1:1 male:female ratio at birth.
Associations between RMSF risk level and numbers of free roaming dogs, dog function, and dog origin were assessed using chi-square tests for rural and urban areas. Fecundity and dog density were compared by risk level (high vs low), and landscape (rural vs urban), using a two-sided Student’s t-test with equal variances. Two-way ANOVA models were performed to evaluate for effects of risk level (high vs low) and landscape (rural vs urban) on means AO, MDM, and NHV of dogs surveyed. Data were analyzed with the statistical program “R” (R-Development Core Team, 2015) with P<0.05 used as a cutoff to infer statistical significance.
Household dog owners
Subscribe to the newsletter news
We hate SPAM and promise to keep your email address safe