Estimating the Dog Population, Responsible Pet Ownership, and Intestinal Parasitism in Dogs in Quito, Ecuador

In 2011, authorities of Quito, the capital city of Ecuador, approved an ordinance to promote public health and animal welfare through responsible pet ownership promotion. The population of dogs was not known, and the relationships between dog abundance, socio-economic factors, prevalence of zoonotic gastrointestinal parasites, and pet ownership responsibility had not been investigated. The objectives of this study were (i) to estimate the human:dog (HD) ratio, (ii) to examine the relationship between household factors and responsible pet ownership and (iii) to estimate the prevalence of households with one or more dogs infected with intestinal parasites in Quito, Ecuador. Space-based random sampling procedures were used for estimation of HD ratios in stray dogs and conned owned dogs. The relationship between household factors and a responsible pet ownership index was examined using logistic regression. Dog fecal samples were tested for intestinal parasites. Among stray dogs, the observed HD ratio was 58:1. Among dogs kept indoors, the observed HD ratio was 3,5:1. A positive interaction effect between number of dogs in study households and household living conditions (a proxy for household wealth) on responsible pet ownership was observed, which we discuss in this report. Prevalence of households with dogs infected with intestinal parasites was 28% (95% CI = 21-37). Ancylostoma spp. was the most frequent intestinal parasite in study dogs kept indoors. This study provides new information that can be used by policy makers to formulate, implement, and evaluate public policies and education programs aimed at enhancing pet ownership responsibility in Ecuador.


Introduction
Responsible pet ownership is a new practice that is becoming more common in the growing, mostly urban Ecuadoran middle class (Castellanos 2016; Universo 2017). However, dog overpopulation, lack of responsible pet ownership, and absence of animal shelter programs are continuing public health issues of concern in Quito, the capital city of Ecuador. In recent years, a number of organizations have advocated for improving pet welfare and in 2011, authorities in the city of Quito approved an ordinance to prevent cruel treatment of animals, promote responsible pet ownership, improve dog health, and reduce the risk of dog bites to people, dog defecation in public spaces, and stray and unwanted animals. In recent years, the number of organizations advocating for the implementation of policies that support responsible pet ownership has increased, as well as the demand for high quality high volume spay neuter (HQHVSN) clinics. A single agent type model utilized in Santa Cruz's (Santa Cruz Island, Galapagos, Ecuador ) dog population recommend an annual spay-neutering between 300/3300 (~12%) to 500/3300 (~15%) of the total population of dogs via HQHVSN clinics to signi cantly reduce the population in 10 years (Diaz et al. 2018). Rural residence, low level of education, and low income have been identi ed as broad factors associated with a lack of responsible pet ownership including normal hygienic practices, responsible breeding, pet adequate nutrition, comfortable housing, mental health, and physical health There are limited published studies related to prevalence of dog with gastrointestinal (GI) parasites in Ecuador and the Andean Region. These studies are useful to provide information on pet and human health conditions. A study in Santa Cruz, San Cristobal and Isabela analyzed the prevalence of intestinal parasitism using a convenience sample of 97 owned dogs that were presented to temporary HQHVSN clinics in Isabela, Santa Cruz and San Cristobal. Fecal samples were collected during a neutering process and processed using the Sheather Sugar otation method. Morphological analysis was used to determine the presence of gastrointestinal parasites. In addition, a commercially available immuno uorescence assay for analysis of Cryptosporidium spp. and Giardia spp. was used. The prevalence of infected dogs for the whole sample population was 71,4 %. The most commonly detected parasites were Ancylostoma caninum (57,7%) and Toxocara canis (16,5%) (Gingrich et al. 2010). Another study in Santa Cruz in 2014 used a randomly selected household sample. A total of 56 fecal samples processed using Sheather sugar otation for morphological analysis stained samples for Cryptosporidium spp. Also, the samples were evaluated by a commercial ELISA for the presence of soluble antigen of Giardia spp. The most prevalent GI parasites in this study were Ancylostoma spp. (38%), Toxocara spp. (5%), and Cryptosporidium spp. (4%). No dogs were positive to Giardia spp ).
Ancylostoma spp. is the most frequent parasite found in studies in the Andean Region (Ramirez- Barrios et al. 2004;Gingrich et al. 2010). The geographic distribution of this parasite is related to the temperature of the location and some species of Ancylostoma spp. are frequently found in places with a temperature higher than 20 o C. Only three species of Ancylostoma are considered zoonotic: A. ceylanicum, A. caninum and A brazilense. A. ceylanicum often successful infecting humans and A. caninum has zoonotic behavior occasionally. On the other hand, A. brazilense is mainly responsible for the "creeping eruption" or cutaneous larva migrans which is the percutaneous infection of L3 in humans. Human infections with Ancylostoma spp. are more common in children and adults that spend time barefoot in areas with warm and moist weather. Professionals with more risk are construction workers and gardeners since their job puts them in direct contact with dirt that could be contaminated with L3. The organ systems involved most commonly are skin, blood and intestine. The infection by skin of the infectious third-stage larvae of A. caninum or A. brazilense will cause skin lesions. Eosinophilic enteritis has been reported in infections with A. caninum. or A. ceylanicum. Infections occur mainly peroral and cause abdominal symptoms (Overgaauw and Van Knapen 2000;Bowman et al. 2010). Toxocara canis is probably the most common GI helminth that infects dogs worldwide. The reported infection rates in domestic dogs vary from 3,5% in adults to 79% in puppies. The mode of transmission in humans is ingesting eggs from contaminated soil, hands and raw vegetables or by the consumption of undercooked meat of a paratenic host. Direct contact with infected dogs is not considered a transmission risk since the parasite ova requires a period of three weeks to embryonate and become infective. In humans, the literature reports that infections in children are more frequent. Severe visceral larva migrans is mainly found in children between 1-3 years of age. Geophagic pica is described as a major risk factor to be infected with T. canis. This behavioral disorder may affect a range of 2 to 10 % of children between 1 and 6 years old (Overgaauw and Van Knapen 2000; Moreiral et al. 2014).
In 2012 Quito authorities requested assistance in estimating the dog population, baseline data on responsible pet ownership and the burden of dogs infected with intestinal parasites in the city of Quito. We present the results of an effort to address these questions by: (i) estimate the HD ratio, (ii) examine the relationship between household factors and responsible pet ownership and (iii) estimate the prevalence of households with one or more dogs infected with intestinal parasites in Quito, Ecuador.

Study site
This study was conducted in the Metropolitan District of Quito, the capital city of Ecuador, during June-August 2013. The estimated human population was about 2,3 million people residing in 634 611 households (INEC 2011a). Quito is divided into 65 Parishes, of which 32 are located in urban areas and 33 in rural areas.

De nitions
Stray dogs were those (owned or not owned) observed stray by one of the authors (CJG) between 4 00 and 6 00 on twoconsecutive days during the study period (World Organisation for Animal Health 2019) Con ned owned dogs were those declared as owned by the household chief and kept indoors or inside household (property) limits at the time of the visit. Some dogs were allowed to roam free (with or without supervision) in public spaces to play or defecate during AM or PM hours.
A household is a housing unit composed by those living together under the same roof (e.g., home or apartment).
The household chief is the person who is in charge of the household at the moment of the data collection.
Commercial food was de ned as food manufactured for the purpose of feeding dogs produced by the industry that generally has a veterinarian specialist in nutrition behind the formulation.
Specially prepared food was de ned as fresh food specially prepared by the dog owner for the dogs in the household.

Sampling and counting of stray dogs
A representative sample of parishes was selected from a map starting with one Parish roughly located in the center. Then each Parish was identi ed with one of four colors (avoiding to assign the same color to adjacent Parishes) ( Figure 1) (WSPA 2008; Hossain et al. 2013). This approach identi ed three groups of 16 Parishes and one group of 17 Parishes (with one color for each group). Second, one of four colors was randomly selected to identify one group of study Parishes. Eight of 16 selected Parishes (Rumipamba, Mariscal Sucre, La Magdalena, La Ecuatoriana, Carcelén, San Isidro del Inca, Puengasí, Solanda) were located in urban areas and eight more (La Merced, Nanegalito, Chavezpamba, Yaruquí, Conocoto, Calderón, Calacalí, Nayón) in rural areas. Third, within each selected Parish, Sections were created using Google Earth®. Each Section had a walking distance of approximately 5 km on public routes and was identi ed with a unique number. Each Parish had a median number of 25 Sections (Range: 3-90). Two Sections were then randomly selected where stray dogs were to be counted using random numbers from a computer algorithm (available at http://www.randomizer.org) Counting of stray dogs was conducted by one author (CJG) on a total of 32 Sections (two selected Sections x 16 Parishes = 32 Sections). Each Section was walked at an average speed of 4 km/h on two consecutive days (e.g., Monday and Tuesday) between 4 00 and 6 00. A metropolitan police o cer accompanied the surveyor for safety. This time period was selected because it is a reported period during which dogs search of food, defecate without dog owner supervision, and avoid morning car tra c (WSPA 2008). A tally counter was used to count dogs each day (e.g., "captured" dogs). An Earthmate ® PN 60 GPS device was used to record the track followed the rst day, and the same track was followed the second day. Dogs that were counted were registered on a digital photographic archive using a Canon® PowerShot D20 camera equipped with a GPS devise. The photographic archive and a second tally counter were used to identify dogs observed on both observation days (e.g., "re-captured" dogs).

Sampling and counting of con ned owned dogs
Within each selected Parish (n=16) and each selected Section (n=32), all Blocks (approximately 14 Blocks per section) were considered for inclusion, but only one Block per Section was randomly selected and included in the study for a total of 64 blocks. All households in selected Blocks were visited for a personal interview with the household chief. Participation in the personal interview was voluntary, and only household chiefs who approved and signed a consent form were included in this study. Blocks in rural Parishes were not clearly de ned. Thus, all households in the selected Section were targeted for inclusion in the study.

Collection of fecal samples and diagnosis of intestinal parasites in con ned owned dogs
During house-to-house visits, each household chief was instructed to collect one fecal sample from at least one of her/his dog(s) from the ground, after normal defecation. On a rst home visit (day 1), the household chief was provided with two pairs of gloves, a disposable tongue depressor and two sterile plastic specimen containers (100 ml) labeled with the dog's name and date. Participant dog owners were instructed how to safely collect dog fecal samples. If canine fecal samples were visible in the home backyard, he/she was provided with zip-lock bags and instructed to safely collect a dog fecal sample from the ground by turning the bag inside out. All fecal samples were collected the following day (day 2) and submitted to a designated laboratory at the University of San Francisco de Quito for identi cation of intestinal parasites. The parasitological exam consisted in identi cation of morphological characteristics of parasite eggs, cysts and oocysts using a Sheather's sugar centrifugational otation and sedimentation techniques ).

Data collection
Poverty rates (%) published in Ecuador's National Population and Housing Census in 2010 were collected for each study parish. For each study household, the following data were collected: date of interview; Parish identi cation; residence (urban, rural); number of people in the household; number of dogs in the household; household with one or more dogs with a positive diagnosis of intestinal parasites (yes, no); responsible pet ownership index (score 1 to 8); and living condition index (score 1 to 9). Finally, for each study dog, the following data were collected: dog name, age, sex (male, female), spay neuter status (yes, no), free-roam (yes, no), dog is walked with a leash (yes, no), adequate shelter (yes, no), feeding (commercial food, specially prepared food, scrap food from human meals), dog visited veterinary o ce in the last 12 months (yes, no), dog was dewormed in the last 6 months (yes, no), and last time dog was vaccinated against rabies (months, don't know).
An instrument was prepared to measure a Responsible Pet Ownership Index (RPOI) in each study household ( Table 1). The instrument included eight parameters (i) spayed/neuter status; (ii) use of leash; (iii) dog shelter conditions; (iv) food; (v) veterinary care; (vi) rabies vaccination compliance; (vii) deworming; and (viii) allowance to free roaming. Scores of 0 to 1 or 0 to 2 were assigned to each parameter for a maximum score of 8 in each household. An instrument was prepared to measure a Living Condition Index (LCI) as a proxy for wealth in each study household ( Table 2). The instrument included two parameters: (i) accumulated wealth (PC, e-mail account, TV, cars) and (ii) education (elementary school, middle school, high school, university). Scores of 0 to 4 were assigned to parameter, except for number of cars, for a maximum score of 9. where n1 is the number of dogs observed during the rst day (captured dogs), n2 is the number of dogs observed the next day, m2 is the number of dogs observed during the rst and second day (e.g., "recaptured" dogs) and N is the estimated number of animals in each section. The rationale for using this method for counting dogs is to estimate the proportion of "recaptured" dogs during the second count (m2/n2); the assumption is that this proportion is the same as that in the population at large (n1/N). Ninety-ve percent con dence intervals (CI) were estimated by using the Poisson distribution because data were not normally distributed, and the number of recaptures was less than 50 in study Parishes (Krebs 1999). Con dence intervals using the two-sample method (Greenwood and Robinson 2006)were only determined for residence (urban, rural) and for the total count since only 4 of 32 study Sections had a required number of ≥ 8 recaptured dogs.
To estimate number of dogs in each selected Parish, the estimated numbers of dogs on Sections 1 and 2 were averaged, then multiplied by the total number of Sections in each selected Parish. To calculate the HD ratio in each study Parish, the o cial human population was divided by the estimated dog population. The sampling fraction method was used to estimate the overall number of dogs by residence in studied Parishes (urban, rural) (WSPA 2008). Overall HD ratios by residence were calculated dividing the o cial human population by the estimated dog population. The overall dog population estimate in all sampled Parishes and the estimated HD ratio for Quito were calculated using the same method.
The relationship between the abundance of stray dogs (HD ratio) (rank data) and o cial poverty rates (rank data) in study Parishes (n = 16) was examined using simple linear regression.

Con ned owned dogs
The observed HD ratio was calculated by dividing the total number of people by the total number of dogs in study households. The variables for number of dogs and number of people in study Parishes (n = 16) were normally distributed. Thus, simple linear regression was used to calculate a beta coe cient to produce an additional HD ratio (i.e., 1 divided by the beta coe cient) and 95% CI derived from the regression results.
The median time of last rabies vaccination and deworming was computed. Wilcoxon sign rank test was used to calculate 95%CI. The frequency of dogs that were examined by a veterinarian within last year was determined.
The association between responsible pet ownership index (RPOI) and investigated household factors was examined by using unconditional logistic regression (Szklo and Nieto 2000; ). Study households were assigned into one of two groups with a low (0 to 5) or high (6 to 8) RPOI scores-based on the median distribution.

Prevalence of and identi cation of exposure factors associated with intestinal parasitism in con ned owned dogs.
Prevalence of households with ≥ 1 dogs with a positive diagnosis of intestinal parasites was calculated as the number of households with a positive diagnosis divided by the total number of households with dogs sampled and tested. In addition, prevalence of dogs with a positive diagnosis of intestinal parasites was calculated as the number of dogs with a positive diagnosis divided by the total number of dogs sampled and tested; 95% CI were calculated for each prevalence estimate assuming normal distribution (Dean et al. 2013). Finally, the relationship between investigated household factors and households with one or more dogs classi ed as infected with intestinal parasites was examined using unconditional logistic regression.

Household factors associated with Responsible Pet Ownership
We found that the responsible pet ownership in a household was in uenced by the residence type, number of dogs, and LCI were all signi cant (p < 0,01). The variable for LCI was associated (p < 0,01) with the variable for residence. Thus, the explanatory variables for number of dogs in a household and LCI were further examined. We found that the combined effect of number of dogs in the household and LCI on RPOI, and the analysis revealed that households with ≥ 2 dogs and low LCI were 16,71 times more likely to have a low RPOI, compared to households with 1 dog and high LCI (OR = 16,71; 95% CI = 5,90-47,37; p < 0,01). This observed combined effect on low RPOI (OR = 16,71) was higher than the expected combined effect based on adding (OR = 7,37), and multiplying (OR = 11,90) absolute independent excesses due to LCI (OR = 2,39) or (OR = 4,98) number of dogs in the household (Table 5) 3.4 Intestinal parasites in con ned owned dogs A total of 110 of 194 (57%) households with one or more dogs returned a dog fecal sample. Thirty-one of 110 or 28% (95% CI= 21-37) study households had one or more dogs classi ed as infected with one or more intestinal parasites. Using univariable logistic regression, the odds of household with dogs infected with intestinal parasites was two times higher in households with low RPOI scores, compared to households with high RPOI scores; however, this association was not signi cant (crude OR = 2,09; 95% CI = 0,90-4,86; p = 0,09). Other investigated household factors were not associated with a positive diagnosis of intestinal parasites (p ≥ 0,12).

Stray dogs
Overall, the estimated HD ratio of stray dogs was 58:1 and the abundance of stray dogs was higher in rural Parishes, compared to urban Parishes. In addition, a higher abundance of dogs was associated with higher poverty rates in study Parishes. These ndings are in agreement with previous observations that a higher abundance of stray dogs is associated with low-or middle-income neighborhoods in Baltimore, Maryland between 1970 and 1971 (Overgaauw and Van Knapen 2000). In another study conducted in rural households in Yucatan, Mexico, 77% of study households did not have adequate fences to prevent dogs from roaming (Beck 2002;Ortega-Pacheco et al. 2007). Although the variable for inadequate fences was not measured in this study, households with inadequate fences or no fences are commonly observed in rural Parishes in Quito. To our knowledge, no other published studies have estimated the abundance of stray dogs relative to the human population in Quito.

Con ned owned dogs
Among 998 selected households, 232 or 23% were surveyed, and the proportion of surveyed households was different between urban (18%) and rural (36%) neighborhoods. The response rate was lower in this study than a survey on Santa Cruz Island, Galapagos in September 2014 (166/227 or 73%) . It is possible that this was due in part to the lack of government o cials participating in the current survey. The prevalence of crime against people in Quito (21,83%) is higher than in the Galapagos (7,06%), consequently, residents may be more reluctant to open the door to a stranger in Quito (INEC  2011b) Among 232 study households, the observed HD ratio was 3,5:1. The observed HD ratio in Quito is smaller than that reported by We observed a positive interaction effect between ≥ 2 dogs in study households and a low household living condition index on responsible pet ownership. This nding can be explained by education and economic factors at the household level, which can in uence compliance for responsible pet ownership as expected by policy makers in Quito, veterinary organizations, and the World Society for the Protection of Animals. Families with a low level of education may not appreciate the dimension of potential health hazards and consequences caused by irresponsible pet ownership practices. Low-income families may not have the nancial means or access to pet veterinary services. Finally, family budgets can be further compromised with a higher number of dogs on the household. In a similar study developed in Pelotas, Brazil, 1558 households were examined for responsible pet ownership. Using similar parameters to develop a score for responsible pet ownership they found an association between level of education of the household chief (p<0,001) and low responsible pet ownership index (Domingues et al. 2015). To our knowledge, no other published studies have examined the relationship between socio-economic drivers and responsible pet ownership.

Intestinal parasites in con ned owned dogs
At the dog level, 25% of dogs were infected with one or more intestinal parasite, with Ancylostoma spp., the most common parasite, identi ed in 12% A. caninum is a parasite that causes cutaneous larva migrans by third stage larvae penetration and migration under the skin of humans; however, in contrast to canids, humans are dead-end hosts (Shalaby et al. 2010). The observed burden of Ancylostoma spp. in dogs in Quito, Ecuador during June-August 2013 was lower, compared to that in dogs on Santa Cruz Island, Galapagos, Ecuador in September 2014 (21/56 or 38%; 95% CI = 25, 50%) ). On Santa Cruz Island, the frequency of dogs infected with Ancylostoma spp. was higher in rural neighborhoods, compared to urban neighborhoods; one explanation offered by  was that environmental conditions (greater exposure and longer survival in soil) are more favorable in rural than in urban neighborhoods in Santa Cruz Island. In this study the prevalence of Ancylostoma spp. was similar in urban households compared to rural households. An explanation for this can be that the samples were taken during the dry season of Quito. Also, there is no major difference on the weather and altitude between the studied sections with the exception of Nanegalito (INAMHI 2011). Ancylostoma spp. is often the most frequent parasite found in dog studies in the Andean Region countries with a prevalences of 24.5% in Maracaibo, Venezuela, 13.9% in Quindio, Colombia, 38% in Santa Cruz and 57.7% in Isabela, Galápagos

Study limitations
This study had several limitations. First, counting of stray dogs was limited to two of a median of 25 sections per selected Parish during two consecutive days, in a narrow time frame between 4 00 and 6 00. It is possible the inclusion of more study sections, additional days of dog counting (e.g., 10 consecutive days), or additional time frames could have produced a higher or lower HD ratio, with more precision.
Second, among con ned owned dogs, the number of surveyed households was 23%. The response rate was higher in rural areas (36%) compared to urban areas (18%) with the consequent reduction of the expected sample size. Third, the RPOI and LCI were estimated based on information provided by household chiefs. It is possible several households could have been misclassi ed as having a low or high Index value. Although in this study, a low RPOI was associated with a low LCI, an indication that observation bias was not high.
Finally, the burden of dogs infected with intestinal parasites was based on one fecal sample collected during the study period.

Conclusions
This study provides new baseline dog population, preventive veterinary medicine practices and veterinary care access data that can be used to formulate, implement, and evaluate the impact of HQHVSN clinics for dogs in Quito, Ecuador. Although the total dog population and their parasite burden was lower than reported in other areas of Ecuador, 28% of households owned dogs with parasites of zoonotic signi cance. Study results support allocation of resources to further promote education programs that can support Responsible Pet Ownership principles (spay-neuter procedures, deworming, access to veterinary services at low cost), particularly in rural neighborhoods, and neighborhoods with high poverty rates.    Figure 1 Geographic location of 65 parishes in the metropolitan area of Quito. Each parish was assigned to 1 of 4 colors (blue, yellow, green orange). Note: The designations employed and the presentation of the material on this map do not imply the expression of any opinion whatsoever on the part of Research Square concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. This map has been provided by the authors.