An analysis of companion animal tick encounters as revealed by photograph‐based crowdsourced data

Abstract Background Community science is increasingly utilized to track important vectors of companion animal disease, providing a scalable, cost‐effective strategy for identifying new foci, changing phenology, and disease prevalence across wide geographies. Objectives We examined photographs of ticks found attached to predominately dogs and cats reported to a photograph‐based tick surveillance program to identify potential areas for improvements in tick prevention education and risk intervention. Methods We compared estimated days of tick attachment using a Kruskal–Wallis one‐way analysis of variance, and a Pearson's chi‐square analysis of variance on the number of submissions by host type submitted for each season. Results The blacklegged tick (Ixodes scapularis) was the most common species reported (39.8%). Tick photographs submitted were almost entirely adults (89.5%), and ticks found on companion animals exhibited an estimated median engorgement time of 2.5 days. Ixodes scapularis displayed the highest median engorgement of the top tick species found feeding on companion animals (χ 2 = 98.96, p < 0.001). Ticks were spotted year‐round; during spring and summer, ticks collected from pets represented 15.4 and 12.8% of all submissions, but increased to 28.5 and 35.2% during autumn and winter, respectively. Conclusions Crowdsourced data reveal that mostly adult ticks are detected on pets, and they are found at a point in the blood‐feeding process that puts pets at heightened risk for disease transmission. The increase in proportion of ticks found on pets during colder months may reveal a critical knowledge gap amongst pet owners regarding seasonal activity of I. scapularis, a vector of Lyme disease, providing an opportunity for prevention‐education.


INTRODUCTION
Tick vectors can pose serious health risks to domestic companion animals and their owners. Dogs, cats and other domestic animals are susceptible to numerous potentially deadly tick-borne bacterial and protozoal diseases, including Lyme disease, spotted fever rickettsioses, American canine hepatozoonosis, cytauxzoonosis and tick paralysis toxicosis (Bowman 2009;Chomel 2011;Kidd & Breitschwerdt 2003;Little et al., 2014;Nagamori & Reichard 2015;Reichard et al., 2010;Shaw et al., 2001). Even in cases where animals are asymptomatic or display sub-clinical signs of tick-borne disease infection, they can act as potential reservoirs for zoonotic pathogens (Fritz 2009;Mather et al., 1994;Shannon et al., 2017;Shaw et al., 2001). Several studies have demonstrated that those living in households with either dogs or cats are at higher risk of tick encounters and tick-borne illnesses than those in households without pets (de Wet et al., 2020;Jones et al., 2002Jones et al., , 2018. Cats that live entirely indoors have been found infested with species of attached ticks that originate in outdoor habitats (Little et al., 2018). This finding suggests that cats, and perhaps even dogs or humans can carry loose ticks indoors that then attach to other hosts. The strong association among pets, their owners and zoonotic tick-borne illness is evident in the literature, as dogs often act as sentinels for tick-borne illness, particularly in cases of Rocky Mountain Spotted Fever (Elchos & Goddard 2003), and Lyme disease (Bowman et al., 2009;Eng et al., 1988;Guerra et al., 2001;Johnson et al., 2004;Lindenmayer et al., 1991;Little et al., 2014;Wagner & Erb 2012).
Recent reports suggest that cases of human tick-borne disease are more than double over the past decade (Rosenberg et al., 2018), no doubt also increasing significantly the tick-borne disease prevention challenges faced by pet owners and veterinary professionals. The factors contributing to more cases of tick-borne diseases are complex.
Broad ecological changes, including alterations in climate and landuse patterns (Allan et al., 2003;Brownstein et al., 2005;Gilliam et al., 2018;Guerra et al., 2002;Hall et al., 2002;Ostfeld & Brunner 2015;Salkeld et al., 2015;Tran & Waller 2013), and increases in deer and other wildlife populations overlapping with human-inhabited space, are highly associated with increases in tick encounters and tick-borne disease prevalence (Ginsberg & Zhioua 1999;Rand et al., 2003;White & Gaff 2018). Additionally, changes in human behaviour resulting in enhanced exposure risks (Fischhoff et al., 2019;Zeimes et al., 2014), and overall improvements in disease diagnosis, surveillance and reporting (Rosenberg et al., 2018) are also responsible for higher occurrence of disease. Concurrently, distributions of tick species are changing, especially blacklegged ticks (Ixodes scapularis Say) and lone star ticks (Amblyomma americanum L.) (Dahlgren et al., 2016;Eisen et al., 2016;Springer et al., 2014), exacerbating the problem while changing climate trends can alter the timing and length of tick and host activity periods, as well as the length of the overlap in companion animal outdoor activity in tick habitat (Alkishe et al., 2021;Ogden & Lindsay 2016;Ogden et al., 2014). Redistribution of wildlife, especially cervids, such as whitetail deer, and movement of ticks on migrating birds, is resulting in the incursion of ticks into more densely populated suburban, peri-urban

IMPACTS
• The blacklegged (deer) tick was the most commonly reported tick attached to companion animals. • Ticks found attached to companion animals have been fed a median of 2.5 days longer than compared to those found attached to humans. The blacklegged (deer) tick had the highest median engorgement of the five top tick species found feeding on companion animals.
• The percentage of reports of ticks found on companion animals is more than doubled in the fall and winter seasons from what was reported in the spring and summer seasons. and even urban areas (Shaw et al., 2001). As a result, new endemic foci (Schwartz et al., 2017;Sonenshine 2018) are putting a larger demographic of humans and pets at risk for tick attachment and illness.
Passive surveillance, particularly involving the contributions of community science, has become an increasingly popular method for tracking ticks on pets and associated pathogens. Often, ticks are collected from pet owners at veterinary offices or other research centres (Johnson et al., 2004;Nieto et al., 2018;Saleh et al., 2019;Xu et al., 2016), identified to species, and tested for a wide range of pathogens. These large-scale collection efforts can be examined over time to observe trends in tick distributions and infection rates (Saleh et al., 2019;Sánchez-Vizcaíno et al., 2016;Tulloch et al., 2017). Ventures using big data, including internet search terms, social media and electronic health records, are also showing promise as a means of tracking tickborne disease in companion animals in an increasingly digital world (Guernier et al., 2016;Tulloch et al., 2019).
Photographs of encountered ticks have been shown to be another reliable method of rapid identification of commonly encountered ticks of medical concern (Fernandez et al., 2019;Koffi et al., 2017;Kopsco et al., 2020) that can be used to not only track tick trends, but also relay important information about potential disease risk to pet owners or veterinarians. The TickSpotters program at the University of Rhode Island's Tick Encounter Resource Centre (TERC) is an online, photograph-based passive surveillance system, staffed by PhD and other graduate-level tick researchers, with an overall accuracy of 96.7% for North American ticks most commonly encountered by humans and their companion animals, particularly I. scapularis, Dermacentor variabilis, and A. americanum (Kopsco et al., 2020). These three tick species are the most commonly encountered tick species by dogs and cats throughout the United States (Duncan et al., 2021;Ghosh et al., 2021;Saleh et al., 2019). The system allows users to submit photographs of ticks along with information surrounding each encounter. Moreover, a response email is sent within 24-h that confirms the identity of the tick to species and stage, estimates a duration window of attachment, and provides tailored guidance regarding likely riskiness for infection and best next actions to prevent disease and avoid future tick encounters.
The purpose of this study was to examine photographs of ticks found attached to pets and reported to TickSpotters photographbased tick surveillance to identify potential areas for improvements in tick prevention education and risk intervention. We anticipated that photographed tick submissions from pets would exhibit unique characteristics that differed from submissions of ticks found on humans.
Specifically, we expected that proportionally more ticks would be detected on pets than on humans during the fall and winter seasons when owners potentially assume that ticks are no longer an active concern, and that ticks attached to pets would be found later in the feeding process (i.e., would display engorgement indicative of a longer feeding duration) than ticks reported from humans.

Data collection
We performed a retrospective analysis on data collected through the  Yeh et al., 1995), and the tick demonstrating the median size was photographed as representative for that time point (Mather, unpublished data). Tick photographs were examined against the TERC pictorial chart displaying different species at half-day fed intervals and the closest estimated feeding window was assigned. Feeding duration of tick species that did not have a representative image was compared to the life stage of the other validated tick species and a 24-h feeding window was estimated. A tailored email response was sent to the participant with species identification confirmation and a riskiness assessment based on the tick/stage-specific diseases and tick infection estimates of those diseases derived from published studies if available for the region where the tick was encountered. Information also was provided in the email on how to prevent future tick encounters and attachments by conducting regular tick checks, using permethrin repellent and various tick 'knock-down' pet products (i.e., products where ticks are not only repelled but by which are also quickly immobilized and killed), and resources to submit ticks for testing should concern exist regarding infection. The collection survey was approved and overseen by the University of Rhode Island Institutional Review Board (# HU1819-092).

Statistics
We calculated basic descriptive statistics and proportions to understand the relative contributions of different tick species and feeding times to the overall submission of ticks from companion animals. To determine whether there were significant differences in duration of attachment (based on an engorgement index) among tick species, we compared estimated days of attachment using a Kruskal-Wallis oneway analysis of variance due to unequal variances and non-normal dis-   to pets for less than 1 day, more than 50% of ticks were attached for approximately 2.5 days or longer (Table 1). The median engorgement of ticks fed on pets was 2.5 days (standard deviation = 2.11, range 1-9 days), while the median engorgement of ticks reported from humans was 1 day (standard deviation = 1.39, range 1-8 days), and those found unattached (either unfed or fed to repletion and detached) also fed for a median of 1 day (standard deviation = 2.29, range = 1-9 days) ( Figure 1). There was a significant difference in tick feeding duration among host type (person, pet, unattached) (χ 2 = 4486.2, p < 0.001), and a large effect size (Cramer's V = 0.28, CI [0.27, 0.29] (Figure 1) (Rea & Parker 1992). All tick feeding duration time categories displayed significantly different (p < 0.001 or 0.01) proportions of hosts with ticks of that engorgement. Feeding duration estimated at the 5.5 day feeding window was not statistically significant because we only received one tick submission that was estimated to be attached for 5.5 days on a human, and at the 9-day estimated time because we received only two ticks estimated at approximately 9 days each (one on a pet and one was found on the floor after detaching from an unknown host) (Figure 1).  Of the five most commonly submitted tick species, blacklegged ticks (both I. scapularis and Ixodes pacificus) displayed the highest median attachment duration at the time they were detected on pets based on their engorgement index (3 days), followed by brown dog ticks (2.5 days), lone star ticks (2 days) and American dog ticks (<1 day).
There was a highly significant difference among these five species (χ 2 = 546.91, p < 0.001) in their attachment duration, but the effect size, or practical magnitude of this difference was only moderate (ε 2 = 0.12, CI [0.10-0.14]) ( Figure 2) (Rea & Parker 1992). Post-hoc pairwise comparisons showed highly significant (p < 0.001) differences between attachment durations at detection for each of these tick species when feeding on pets ( Figure 2).
The proportion of ticks submitted from pets was significantly different from those found on humans and from tick found unattached in each season (p < 0.001), as well as across seasons (p < 0.001 (Figure 3). In

DISCUSSION
We described 5132 domestic companion animal tick encounters reported to a photograph-based passive tick surveillance system during a 4-year period. Overall, we found that this group of domestic animals encounters commonly occurring tick species that present a potential disease risk, and also pose a threat to the humans with F I G U R E 2 Violin boxplot displaying engorgement at detection (days) of the five most commonly reported tick species found on pets that were submitted to TickSpotters from 2014 to 2018. Feeding times were estimated by comparison to a pictorial tick engorgement chart created by TERC that removed feeding ticks from hosts at half day intervals, and was based on the scutal index. The horizontal bar within the box represents the median engorgement for each species. The dots represent the mean engorgement. Kruskal-Wallis nonparametric analysis of variance tested statistical difference among the species engorgement distributions. Effect size is denoted by episilon 2 . Post-hoc pairwise comparisons were conducted using the Dwass-Steel-Crichtlow-Fligner method and (Holm) adjusted p-values for statistically significant distributions are noted between species whose overall submissions were statistically different in engorgement whom they share a household (Jones et al., 2018). Blacklegged ticks (I. scapularis) were the most frequently reported species encountered by companion animals, and the species demonstrating the highest median feeding duration rate of the five most commonly encountered ticks. It is well documented that risk for transmission of tick-borne pathogens increases with longer duration of tick feeding (des Vignes et al., 2001;Dolan et al., 2017;Eisen 2018;Kidd & Breitschwerdt 2003;Piesman et al., 1987;Sood et al., 1997). The proportion of ticks reported from pets was more than double in the fall and winter months ( Figure 3) indicating that there remains a general lack of awareness that colder months still pose a risk for tick encounters. These findings echo similar results from other investigations into ticks reported from pets (Saleh et al., 2019). Taken together, these observations are troubling due to the numerous diseases that I. scapularis can vector and transmit to both pets and humans. A recent survey of pet owners found that people whose pets had a tick encounter within the past 6 months were more likely to encounter a tick themselves (de Wet et al., 2020). Opportunities exist to increase awareness on this point among pet owners through education by both veterinarians and public health agencies promoting a One Health focus to tick-borne disease prevention. We propose below several targeted areas for educational intervention that can potentially address our findings and help protect pet and human health.
It is imperative that the public have knowledge of seasonal tick activity in their locale (Dryden & Payne 2004). Misconceptions regarding which ticks are capable of transmitting which diseases is pervasive (Halperin et al., 2013) and can lead to both undiagnosed disease cases, and falsely believing that one is infected. In particular, Lyme disease is often surrounded in controversy and confusion (Auwaerter et al., 2011). Awareness that I. scapularis and I. pacificus ticks potentially carrying the Lyme disease agent are still active in the fall and winter would help support year-round tick prevention in areas where these ticks occur (Dryden & Payne 2004). Our findings emphasize the importance of more broadly publicizing these ticks' seasonal activity.

F I G U R E 3
Comparison of seasonal tick encounter submissions by host (i.e., on whom or how the tick was found) out of the total United States submissions (no pet tick encounters were submitted from other countries) using Pearson's chi-square analysis. Effect size is denoted by Cramer's V. ***p < 0.001 We recommend that this information be disseminated before and during the fall tick season via veterinary clinics, mass media and through public health social media pages.
Tick checks on pets are important for detecting and removing loose and wandering ticks before attachment, and early on in the feeding process due to transmission delays for pathogens (Dolan et al., 2017;Eisen 2018;Kidd & Breitschwerdt 2003;Piesman et al., 1987;Sood et al., 1997). Pet checks should be conducted routinely, both immediately after outdoor activity and during regular intervals after the pet has come indoors, to catch any ticks that may have been missed in previous checks. Common areas for tick attachment include the head and ears, back, axillary, inguinal, feet and tail/perianal regions (Little et al., 2019;Saleh et al., 2019), and these ticks should be removed as soon as they are found using fine-tipped tweezers . While it may be that the higher proportion of 'found' ticks on pets than humans during fall and winter is related to fewer ticks successfully attaching to humans due to more clothing barriers, the longer duration of tick attachment on pets compared to humans is suggestive that pets are not being checked regularly or thoroughly enough, or that owners are less inclined to continue use of effective tick prevention products when they believe tick activity is lower.
Effective industry-tested products, including topical tick preventative (synthetic pyrethroid-based formulations), systemic medications (isoxazolines) and vaccines (e.g., against Lyme disease-causing bacterial species) are a necessary part of the tick-borne disease prevention toolkit in pets (de la Fuente et al., 2015;Littman et al., 2018). Numerous topical products exist but often do not repel or kill all species of ticks, and many require a tick to attach in order to get a lethal dose of the pesticide. We first recommend products that repel and kill or immobilize ticks before attachment, but suggest veterinarians and public health professionals discuss with pet owners their concerns regarding pesticide usages and develop a solution that both placates fears regarding synthetic chemical usage and still keep pets protected (Isman & Grieneisen 2014;Johnson 2018 In most cases where less commonly encountered ticks were suspected, we did not request that the participant submit an in-hand specimen for identification unless the suspected tick seemed out of its expected geographical range or seasonal activity. One exception to this was for suspected H. longicornis due to its close similarity to Haemaphysalis leporispalustris and the currently standard procedure for confirming this tick by molecular diagnostics or submission to the USDA National Veterinary Services laboratory. Whenever a Haemaphysalis spp. tick photograph was submitted, we sent a prepaid envelope to the participant to mail us the specimen so that it could be confirmed by microscopic investigation, and then by molecular sequencing at the USDA National Veterinary Services Laboratory. The method used to estimate tick feeding duration based on the scutal index was only validated for three species and their life stages under laboratory conditions, so we were only able to provide a window of attachment estimates, instead of precise time points. Ticks feeding times can vary to some extent depending on various conditions, including temperature (Pollock et al., 2015) and host/tick infection status (Couret et al., 2017), but a 24-h window of feeding time provides appropriate coverage of most tick-borne pathogen transmission windows. However, this estimate may not be sufficient for properly identifying risk if a tick was subjected to interrupted feeding, and pathogen transmission time was expedited (Tahir et al., 2020). We recognize that the north-eastern and mid-Atlantic regions of the country were overrepresented in this sample, and therefore, these results cannot be extrapolated to all areas of the United States, and future work should strive to obtain a balanced proportion of submissions from all regions.
In cases where there was a question about photograph quality, additional photographs were requested of both ventral and dorsal sides of the tick. Since the submission of samples was heavily skewed to the northeast and middle Atlantic regions of the United States, the description of which tick most frequently infests companion animals cannot be generalized to, and may not reflect the larger population of pet encounters. Additionally, many of the 3,030 ticks classified as loose and wandering in the TickSpotters surveillance system were fully engorged ticks that could have fallen off of a pet. Since we did not know for certain if there was a pet in the household in those cases, we still provided pet-tailored prevention information. Ultimately, our investigation identified several important opportunities for prevention-education to pet owners, and suggests the utility of this type of surveillance for companion animal tick encounters.

ACKNOWLEDGMENTS
This work was supported, in part, by the USDA National Institute of Food and Agriculture, URI Cooperative Extension and the RI Agricultural Experiment Station. We also thank Dr. Marie Varloud for helpful formative discussions, and CEVA Animal Health for partial financial support. Thank you to the reviewers and editors for their comments and suggestions that improved this work.