Health indicators in surplus calves at the time of arrival at auction markets: Associations with distance from farms of origin in Québec, Canada

This study evaluated the associations between estimated distance from farms’ locations to auction markets, and health indicators of surplus dairy calves sold during summer 2019 and winter 2020 in Québec, Canada. A total of 3,610 animals from 1,331 different farms were used in this cross-sectional cohort study. Geographic coordinates (latitude and longitude) were obtained for each farm and the 2 participating live-stock auction markets. Calves’ abnormal physical signs (APS) were noted upon arrival at the auction market as they were examined by trained research staff. The haversine distance between the farm and the auction market was evaluated using geographic coordinates and categorized. Generalized linear mixed models were used for statistical analyses. The main APS observed were ocular discharge (34.9%), abnormal hide cleanliness (21.2%), swollen navel (17.2%), dehydration score 1 (at least one of the 2 following clinical signs: persistent skin tent or sunken eye, 12.9%), and dehydration score 2 (both clinical signs mentioned above, 6.5%). Calves from farms located at greater distances from the auction markets (≥110 km) had a higher risk ratio [RR = 1.08; 95% confidence internal (CI) = 1.03, 1.13] for dehydration than those from lesser distances (0–25 km). During the summertime, a RR of 1.18 (95% CI = 1.15, 1.22) was observed for dehydration compared with wintertime. A 2-way interaction between estimated distance and season showed a higher prevalence of ocular discharge for calves from farms at distances greater than or equal to 110 km during the summer (RR = 1.11; 95% CI: 1.04, 1.20) than for calves from farms located at lesser distances (0–25 km). These results demonstrate that calves from farms located at greater distances from the auction markets had more APS, mainly during the summer. A better understanding of the transport conditions and interaction with management at the farm of origin is determinant to mitigate the impact of the journey on surplus calf health.


INTRODUCTION
A significant number of calves from dairy farms in Canada are sold at an early age for different purposes including veal (Wilson et al., 2020a) or dairy-beef crossbred (Renaud et al., 2017) production.In the Canadian province of Québec, surplus calves from dairy farms are commingled (generally at less than 2-3 wk old) in auction markets and bought by owners of veal units.The calves are therefore transported from their farms of origin to the auction markets over various distances (Wilson et al., 2020b;Buczinski et al., 2021).In a retrospective study, using information covering 12 yr of auction market sales in Québec, the median distance between the farm and the auction market was estimated to be 52 km (interquartile range 30-95 km; Buczinski et al., 2021).
Distance was associated with the price of calves sold at Québec's auction markets (Buczinski et al., 2021), and in another study, in the same region, unhealthy characteristics (based on veterinarian diagnosis) were also associated with the price (Marquou et al., 2019).They found that 20.3% of calves had omphalitis (swelling, discharge, or pain), 7.7% had unhealthy characteristics, 6.3% had diarrhea, and 2.5% had nasal discharge at different auction markets.In an Ontarian study, clinical alterations upon arrival at the veal facility were also described; approximately 13% of calves had diarrhea, 25% had an abnormal umbilical score, and 47% were dehydrated (Renaud et al., 2018a).On the day of arrival at grower in Ohio, 35.1% of the calves were dehydrated using the skin-tent test, 27% had inflamed navel, 14% had diarrhea, and 14% were depressed (Pempek et al., 2017).In an Australian study evaluating young calves transported to abattoirs, 12% of the calves had signs of anemia and 11% had a high concentration of urea, compatible with dehydration (Roadknight et al., 2021).
Young calf transportation can be a risk factor for the occurrence of health problems (Roadknight et al., 2021).Dehydration, based on packed cell volume assessment, was associated with long-distance travels in Australian bobby calves (Roadknight et al., 2021).Abnormal umbilicus, dehydration, and sunken flank were associated with early mortality (Renaud et al., 2018b), and each additional hour of travel increased the odds (odds ratio = 1.45; 95% CI = 1.18, 1.76) of calf mortality (Boulton et al., 2020).Studies on the relationship between clinical signs and distance from farms to auction markets are relatively scarce in Québec and North America.The high variation of weather conditions between winter and summer could also impact the effect of transportation on calf health, as previously shown in feedlot cattle (Cernicchiaro et al., 2012).
Our objective was therefore to evaluate the association between distance from the farms to the auction markets and health characteristics of surplus dairy calves in Québec.Our hypothesis was that traveling long distances is negatively associated with various indicators of calf health, and the effect may depend on season and weight.Breed, auction market, and sex can also be related to health characteristics.

MATERIALS AND METHODS
This cross-sectional cohort study was conducted between June 2019 and February 2020 in the province of Québec, Canada, using data collected twice per season on the 2 largest livestock auction markets.The study was performed in accordance with the local animal use ethical committee (protocol number 20-Rech-2058).

Calf Health Indicators and Characteristics
Calf health indicators (presence of dehydration, ocular discharge, nasal secretion, limb aspect, abnormal hide cleanliness, ear carriage, umbilical characteristics, such as wet, swollen, or painful navel, and poor body condition) and characteristics (breed and sex) were assessed by trained research staff upon arrival at the auction market, as previously described (Buczinski et al., 2022).Dehydration was evaluated based on the presence of sunken eyes and skin elasticity on the neck after a skin-tent test (<2 s, ≥2 s).Calves were classified as not having any clinical sign of dehydration (score 0), calves with one clinical sign of dehydration (score 1), and calves with both clinical signs (score 2).Presence of ocular discharge (purulent discharge or crusts or serous epiphora vs. absence), and nasal secretion (presence of purulent discharge vs. absence) were also noted.Another health indicator included was an abnormal limb aspect (lameness, abnormal joints, or limb wounds).Hide cleanliness was evaluated using the absence (clean) or presence (dirty) of liquid (recent diarrhea) or dry fecal material (indicator of previous diarrhea or dirtiness) in the perineal area.Presence of abnormal ear carriage was also assessed.The umbilical stump was evaluated for the following characteristics: wet umbilical area, swelling (more than 3 cm width or 2-fingers, thickness), and pain at palpation (evidence of abdominal contraction or manifestation of pain with the calf rapidly moving his legs or kicking).Poor body condition was also assessed by visual inspection.A calf was diagnosed as thin if it appeared emaciated with little muscle or fat and clearly defined bone structure, vertebrae, and multiple ribs seen similar to a recent study on bob veal calves focusing on animal with poor body condition (England et al., 2023).In a previous study, using the same database, the reliability of these clinical signs was published (Buczinski et al., 2022).A total of 19 calves were evaluated by 8 raters and a weak Gwet agreement type 1 coefficient (AC1) was observed for ear carriage (AC1 = 0.3), and a strong agreement (median AC1 for the all assessed signs: 0.95, range 0.87-1) was observed for the other indicators (Buczinski et al., 2022).The calves' BW (kg) were obtained from the official auction scales.The breed phenotype (dairy or dairy-beef crossbred) and sex (male or female) information were systematically collected by the auction system and also recorded.

Postal Codes and Environmental Temperature
The farm's postal codes were obtained at the auction markets during the sampling period, and farm's location (latitude and longitude) was obtained using a free database in Service Objects Insights on Demand (Santa Barbara, CA; https: / / www .serviceobjects.com/address -geocode/ ).Data validation was performed for latitude and longitude.A total of 8 farms from the main database were selected randomly according to farm's postal codes.A new database with these selected farms was developed selecting the postal codes, latitude, and longitude of the farms.New coordinates were obtained using the farms' postal codes using Google Maps.These data were included as new columns in the new database.The association between the latitude obtained from the free database Service Objects Insights on Demand and the latitude from Google Maps was evaluated.The same was performed for longitude.Monthly environmental temperatures (lowest minimum, mean, and highest maximum) and Québec's stations' coordinates (latitude

Statistical Analysis
Data analyses were performed using R software (R Development Core Team, 2022).The hypothesis of this study was illustrated with a directed acyclic graph (DAG) using the web application DAGitty (Textor et al., 2016).The DAG has been proposed in different studies to explain possible direct and confounding effects (Suttorp et al., 2015;Textor et al., 2016).Interactions between predictors (distance × season and distance × weight) on the different outcomes (presence of dehydration, ocular discharge, nasal secretion, limb aspect, abnormal hide cleanliness, ear carriage, umbilical characteristics, such as wet, swollen, or painful navel, and poor body condition) were our main hypotheses in this study, and traditional DAG has a limitation in these cases (Weinberg, 2007;Nilsson et al., 2021).An adjusted DAG was developed to clarify the study hypothesis (Figure 1).Plausible confounders such as presence of other clinical signs in the different models (mostly dehydration signs) were also tested but discarded if not changing the coefficient <20%.
The quantification of the estimated distance between farms and auction markets was calculated using the same methodology described by Buczinski et al., (2021) using the package "geosphere" (Hijmans et al., 2021).In this package, the haversine function was used to obtain the distance (d) between different positions' coordinates, using the latitude and longitude of each data point (Equation [1]): where r is the radius of the sphere (earth) and latitude (lat) and longitude (long) are the coordinates of the 2 different points [site of origin (1) and auction market (2)].
Using another approach, considering the best route, estimated distance, and duration of travel (minimum distance and time between farm's location and auction markets, respectively) were obtained using the osrmroute () function from the "osrm" R package, assessing "Open Source Routing Machine" (OSRM), based on OpenStreetMap data.Minimum distance and time were referred because this is an estimation of 2 coordinates' points, and the events during the trip were not evaluated in this study.As data validation, the association between estimated distances obtained using the haversine distance and estimated distances by OSRM API was evaluated.
The haversine distance between Québec's different meteorological stations and auction markets was also calculated, and stations located at distances greater than the maximum distance between farms and auction markets were excluded.June and July 2019 monthly temperatures were used to access the average for the summer, while February 2020 monthly temperatures were used for the winter, according to the data collection period.Interpolation models were developed accessing the coordinates from different meteorological stations in Québec, and proximity polygon graphics were reported.We adopted "nearest neighbor" interpolation as the methodology, using the "rspatial" R package.All polygon graphics were rasterized using the function rasterize () from the "raster" R package.
Descriptive statistics were obtained for each predictor (distance, season, weight, breed, auction market, and sex), temperature, and health indicators (presence of dehydration, ocular discharge, nasal secretion, limb aspect, abnormal hide cleanliness, ear carriage, umbilical characteristics, such as wet, swollen, or painful navel, and poor body condition).Generalized linear mixed models (GLMM) using logistic links were used to study the relationship between predictors and outcomes.Distance, season, weight, breed, auction markets, sex, and 2-way interactions between season and distance, and weight and distance were offered as a predictor (independent variables).Abnormal physical signs (APS) were defined as outcomes (dependent variables).A specific model was developed for each APS: where Y = (2, 1, 0) was used for dehydration, and Y = (1, 0) for dichotomous outcomes.The constant intercept was represented by α.X i represents the independent variables and X j × X j′ the relevant interactions, with regression coefficients represented by the vector β.
To adjust for herd clustering, the site of origin of the calves (farm) was defined as a random intercept term u, which was assumed as following a normal distribution (mean = 0, variance = σ 2 ).P = probability of the outcome occurring, n = number of independent variables, i = independent variables, k = number of interactions, j = relevant interactions.Two strategies were used to avoid collinearity problems.First, with the development of a DAG, and second, with the Spearman rank correlation using a pairwise approach to assess the association between predictors.Important predictors were kept in the multivariable model using a liberal P-value of 0.2 (Dohoo et al., 2010).Modeling was performed using a manual backward stepwise strategy until all remaining variables were associated with the outcome using a Wald chi-squared test.Interaction terms were assessed and included in the final model only if the association with the outcome was significant (P < 0.05; Dohoo et al., 2010).
A final model was developed for each APS (dehydration, limb aspect, abnormal hide cleanliness, ear carriage, umbilical characteristics such as wet, swollen, or painful navel, and poor body condition).The fit of the models was assessed with the Nagelkerke pseudo R 2 (ordinal models), marginal and conditional R 2 (binomial models), and Akaike's information criteria.Ordinal regressions were developed using the polr () function from the "MASS" R package (Venables and Ripley, 2002).Binomial GLMM was performed using the lme4 R package (Bates et al., 2015).Regression coefficients, odds ratio (OR), confidence intervals (95% CI), and P-values were obtained by gtsummary (Sjoberg et al., 2021), and reported.Statistical significance was set at P < 0.05.
Risk ratio is a parameter recommended in situations that the outcome is common and when we have a cohort study design (Zou, 2004;Hudson et al., 2005).Because of that, to estimate this parameter with a robust error variance, another GLMM model was developed for each selected outcome and significative predictors (previous ordinal and binomial models) using the Zou's modified Poisson regression (Zou, 2004).Outcomes were considered as count data due the selection of the Poisson (link = "log") family.Sandwich error estimator was applied for each model.Regression coefficients were obtained using the same strategy used for ordinal and binomial models, however, to obtain risk ratios and their confidence intervals, the modified methodology (Zou, 2004) was used.
Graph effect displays, specific to complex GLMM, with the binomial and ordinal models, were developed using the "effects" and "sjPlots" R packages to present the predicted probabilities (marginal effects) corresponding to each level of response according to significant predictors, direct effect, and interaction (Fox, 2003;Lüdecke, 2018).

General Description
A total of 3,656 animals from 1,349 different farms, with a median of 2 calves per farm [first quartile (Q1) = 1, and third quartile (Q3) = 4] were used in this study.Some farms' latitude and longitude coordinates could not be identified in the free database (Service Objects Insights on Demand), and after data cleaning, a total of 3,610 animals from 1,331 farms were kept for statistical analysis.A strong association was obtained for the coordinates from this database and Google Maps (Pearson correlation coefficient; latitude = 0.98, and longitude = 0.99).

DISCUSSION
In this study, we specifically focused on the association between estimated distance between farms and auction markets and the presence of APS in surplus dairy calves.Estimated distance was associated with various physical signs including dehydration, presence of ocular discharge, and abnormal hide cleanliness.As expected, the risk of dehydration increased with distance, especially during the summer.Dehydration was described as an important issue during transportation of young calves (Renaud et al., 2018b;Roadknight et al., 2021) and an important risk factor for mortality (Boulton et al., 2020).Dehydration can be caused by various conditions that could potentially be worsened by specific transportation conditions such as duration and management before and during the journey (Knowles et al., 1997;Cernicchiaro et al., 2012;Marcato et al., 2020), and heat stress (Burhans et al., 2022).
According to new provisions in Part XII of the Health of Animals Regulations from the Canadian Food Inspection Agency (CFIA, 2020), preweaning calves can be transported for a maximum of 12 h without feed, water, and rest.In our study, which started before the application of this regulation, using OSRM, a median time of 1.26 h (Min = 0.02 h, and Max = 10.6 h) for time estimation between farms and auction markets was obtained.This estimate does not account for possible stops of the trucks, time of stay in the auction, and transport to the feeding unit.The management before and during transportation is another important factor associated with the occurrence of dehydration (Knowles   , 1997).In one study performed in Canada, dairy calves did not receive milk or water before or during transport for an estimated period of 24 to 36 h (Wilson et al., 2020b).In another study from Québec, 9.6% of farms did not provide access to water for preweaning calves (Vasseur et al., 2010).Water restriction is an important cause of dehydration, mainly during heat stress (Burhans et al., 2022).
Heat is another important factor that can explain the dehydration observed during the summer in our study.High monthly temperatures were reported during the summer in a large portion of Québec (Supplemental File S2).Dehydration in heat stress conditions occurs due to body fluid losses (Fisher et al., 2014), associated with thermoregulation.Calves suffering from heat stress can present alterations such as tachypnea, hyperventilation, ptyalism, acid/base, and electrolyte imbalances (Burhans et al., 2022).In extreme cases, death may occur due to shock or coagulation problems.The effect of a high ambient temperature on water intake by young calves was described in a previous report (Quigley, 2001).These calves, maintained in low temperatures (0°C) had a predicted water intake approximately of 1.4 L/d, and in high temperatures (>30°C) more than 3 L/d was observed in addition to milk ingestion (Quigley, 2001).
The fact that the calves presented dehydration in the summer could also be associated with the type of truck and ventilation.In a German study, calves transported in open trucks had lower sodium concentration values compared with calves transported in conditioned trucks (i.e., trucks with ventilator system and climate control; Marcato et al., 2020).Sodium losses occur in animals under heat stress (Burhans et al., 2022).In open trucks, the calves are vulnerable to the climate changes (i.e., heat stress).The ventilation needs to be adjusted according to the temperature changes, and an adjustable weather panel is a recommended system (National Farm Animal Care Council, 2001).No studies were found regarding the ventilation control in trucks used for the transportation of calves in different seasons, but it can be expected that truck ventilation is not the same between seasons.The type of trucks was not evaluated in our study; however, it can be an important cofounder for the presence of APS.
Distance from farms to auction markets was also associated with the presence of ocular discharge.Despite not assessing respiratory disease per se during the physical examination, the prevalence of other indicators of respiratory diseases such as nasal secretion or abnormal ear position (associated with pneumonia/otitis complex) were low.A low prevalence of respiratory disease in young calves (less than 1 mo) has previously been reported (Dachrodt et al., 2021;Denis-Robichaud et al., 2021).Respiratory disease is also more commonly observed during the winter period, whereas ocular discharge is more common in calves examined during summer auction sales.Therefore, lacrimal stimulation by any bovine respiratory disease agent is unlikely.We hypothesized that the presence of ocular discharge could be partly related to specific environmental challenges (i.e., contact with particles coming from the environment, asphalt or vehicles, and ventilation), as more specifically observed for long-distance transportation and during summer.In hot or humid weather, some points must be considered, such as ventilation  adjustment according to the weather, transportation at night or early morning, and avoiding periods of intense traffic (National Farm Animal Care Council, 2001).Interestingly, calves transported for long distances had fewer cases of abnormal hide cleanliness (as an indirect indicator of diarrhea), and this may reinforce that the dehydration cases observed in long distances were more related to factors other than diarrhea.No association between watery feces and transportation was observed in another study using a limited number of calves (Uetake et al., 2011).Calves transported for long distances (i.e., coming from farms that were far away from the auction markets) may be prepared or selected for transportation differently from farms selling calves closer to auction market.This could be a confounder for this finding on abnormal hide cleanliness.Driver characteristics (i.e., experience or driving attitude) may also be different depending on transportation distance and this could justify further investigations.In the summer, fewer cases of abnormal hide cleanliness were observed.Management during transportation can be different during the different seasons.In Québec, during the summer, cleaning trucks and facilities could be easier than during winter because access to water is easier.For example, in winter for trucks not having an indoor heated place to be washed and dried, it could be anticipated that cleaning outside would be challenging due to the risk of freeze of water line and risk of freeze of remaining water in the truck before complete drying.
Dairy-beef crossbred calves had a lower prevalence of dehydration, ocular discharge, and poor body condition than dairy calves.The lower body fat reserves of dairy calves could explain the higher prevalence of dehydration and poor body condition during transportation (Fisher et al., 2014;Roadknight et al., 2021).Animals transported to site B had more APS than at site A. Variation in the occurrence of clinical anomalies in different Québec auction markets was previously observed (Marquou et al., 2019), as well as for general calf prices, probably due to difference of calf buyers at different auctions (Buczinski et al., 2021).Finally, female calves generally had less APS than male calves.Deficiencies in handling male calves is recognized in Canada (Renaud et al., 2017;Wilson et al., 2020a), emphasizing the need to improve management for these animals.
This study was developed using a database developed for another purpose, and for that reason, the sample size estimation was not performed previously.However, after data collection, a simple approach based on logistic regression approach was performed (Peduz et al., 1996) in the absence of specific recommendations for Zou's modified Poisson regression.In this approach, the sample size is the result of the multiplication of a factor (10) by the number of predictors (independent variables) divided by the prevalence of a clinical event.
In Table 2, it is possible to see a large variation for the prevalence of APS.The presence of nasal secretion, for example, was the lowest prevalence (0.6%), and the presence of ocular discharge was the highest (34.9%).If we use 6 (number of predictors) multiplied by the factor (10) and divide that by 2% (prevalence of an APS), a sample size of 3,000 will be necessary.For our study, most of the prevalence of APS was higher than 2%, and with a total sample of 3,610, we think our sample size was adequate.
Also, it is important to note that 2 different statistical models (generalized linear models using logistic and Poisson family) were applied to evaluate the associations between predictors and outcomes.A similar result was observed comparing the coefficients regression and P-values from the logistic and modified Poisson regressions.However, as a relevant find for the quality of the results, the values of RR from the modified Poisson regression were lower than the OR, controlling the overestimation.As an example, for the presence of dehydration, calves transported at summertime had an OR of 3.03 and a RR of 1.18 compared with wintertime (Supplemental File S4, https: / / figshare .com/articles/     Ramos, 2023).This problem was foreseen (Zou, 2004;Hudson et al., 2005) and it can justify the use of RR in our study.
Some limitations of this study need to be highlighted.The estimated distance was obtained using the haversine formula (Hijmans et al., 2021).This is not an indicator of the exact distance traveled.For example, there can be differences between this straight line and the road distance traveled.Moreover, even if a road distance could have been used, it is still an approximation that does not account for stops during transportation, number of stops, time, precise itinerary, and driver behavior used for collection of calves during the journey.The study was performed using punctual assessment of the sales, and the associations found need to be confirmed by future studies.In addition, commercial transport conditions are not homogeneous, increasing the risk of unmeasured confounders.Commercial surplus calves' transportation is not extensively studied in North America, and this observational study is an important step to determine the potential effects of new transportation regulations and can serve as a preregulation observation of surplus calf characteristics in Québec auction markets.

CONCLUSIONS
This is the first study reporting the association between estimated distance from farms to auction markets and health of surplus calves in Québec, Canada.Several physical alterations were observed in calves at the auction markets and depended, among other factors, on the distance from farms and season of transport.We highlight the need for interventions and new studies on calves coming mainly from long-distance farms and during the summer to improve calf health at auction markets in Québec.A specific insight on the remainder of the calf journey (waiting period in the auction market and transport to the dairy-beef unit) should also be considered.
Figure1.Directed acyclic graph to evaluate the relationship between exposure variables and outcomes.Abnormal physical signs (APS; i.e., the presence of dehydration, ocular discharge, nasal secretion, limb aspect, abnormal hide cleanliness, ear carriage, umbilical characteristics, such as wet, swollen, or painful navel, and poor body condition) can be influenced by distance (km) and its interactions (arrows directed to a line connecting two variables) with season (summer vs. winter) and weight (kg).The predictors breed, auction market (A and B), and sex (male and female) can directly affect APS.Distance was the estimated distance in kilometers (km) between the farms and auction market; weight was obtained from the auction market; auction represents 2 different large livestock auction markets in Québec, Canada.

Figure 3 .
Figure 3. Predicted probabilities (marginal effects and 95% CI) of a multivariable model (interaction between distance and season) for explaining ocular discharge in 3,610 calves sold at 2 livestock auction markets in Québec, Canada, between June 2019 and February 2020.

Figure 4 .
Figure 4. Predicted probabilities (marginal effects and 95% CI) of different multivariable models for explaining ocular discharge (A, B, C) and abnormal hide cleanliness (D, E, F) in 3,610 calves sold at 2 livestock auction markets in Québec, Canada, between June 2019 and February 2020.

Figure 5 .
Figure 5. Predicted probabilities (marginal effects and 95% CI) of different multivariable models for explaining wet (A, B, C, D) and swollen navel (E, F) of 3,610 calves sold at 2 livestock auction markets in Québec, Canada, between June 2019 and February 2020.
figure/ Ramos _SFile _explanation _pdf/ 22723615/ 3;Ramos, 2023).This problem was foreseen(Zou, 2004;Hudson et al., 2005) and it can justify the use of RR in our study.Some limitations of this study need to be highlighted.The estimated distance was obtained using the haversine formula(Hijmans et al., 2021).This is not an indicator of the exact distance traveled.For example, there can be differences between this straight line and the road distance traveled.Moreover, even if a road distance could have been used, it is still an

Figure 6 .
Figure 6.Predicted probabilities (marginal effects and 95% CI) of multivariable model for explaining poor body condition of 3,610 calves sold at 2 livestock auction markets in Québec, Canada, between June 2019 and February 2020.
Ramos et al.: HEALTH INDICATORS IN SURPLUS CALVES

Table 1 .
Ramos et al.: HEALTH INDICATORS IN SURPLUS CALVES Description of predictors of 3,610 calves sold in 2 livestock auction markets in Québec, Canada, between June 2019 and February 2020 2Letters A and B refer to 2 different auction markets located in Québec, Canada.

Table 2 .
Ramos et al.: HEALTH INDICATORS IN SURPLUS CALVES Description of abnormal physical signs of 3,610 calves sold in 2 livestock auction markets in Québec, Canada, between June 2019 and February 2020

Table 3 .
Generalized linear models(Zou's modified Poisson regressions)to evaluate the association between independent variables and abnormal physical signs in 3,610 calves sold in 2 livestock auction markets in Québec, Canada, between June 2019 and February 2020 1 Ramos et al.: HEALTH INDICATORS IN SURPLUS CALVESIndependent variables include distance, season, weight, breed, auction markets, and sex; abnormal physical signs include dehydration, ocular discharge, abnormal hide cleanliness, umbilical characteristics (such as wet, swollen, or painful navel), and poor body condition.Binomial and ordinal results with the odds ratios are reported in Supplemental File S4 (https: / / figshare .com/articles/figure/ Ramos _SFile _explanation _pdf/ 22723615/ 3; Ramos, 2023).Description of the results according to each dependent variable (abnormal physical sign).

Table 3 (
Continued).Generalized linear models(Zou's modified Poisson regressions)to evaluate the association between independent variables and abnormal physical signs in 3,610 calves sold in 2 livestock auction markets in Québec, Canada, between June 2019 and February 2020 1