Behavioural traits to identify subclinical diseases of grazing ruminants : A thesis submitted in partial fulfilment of the requirements for the Degree of Doctor of Philosophy at Lincoln University

Abstract

Animal welfare and wellbeing can be improved through the early detection and diagnosis of disease. Subclinical infections in particular can be difficult to determine, however, animals may demonstrate subtle changes in behaviour. These changes related to eating, ruminating, inactive behaviour and active behaviour can potentially be identified and measured through sensor technologies. This thesis composes of a series of studies designed to explore changes in behavioural patterns of grazing ruminants experimentally challenged with subclinical infections using commercially available accelerometers. The objective of the first trial (Chapter 3) was to investigate the changes in grazing, ruminating and activities of grazing dairy calves infected with the internal gastrointestinal nematodes Ostertagia ostertagi and Cooperia oncophora and determine the behavioural patterns after receiving an anthelmintic treatment, using neck mounted sensor collars. At Day -12 relative to an experimental infection, animals were dewormed and fitted with Allflex EveryCow collars which can identify mutually exclusive behaviours such as eating, ruminating, resting, walking, medium activity, high activity and behaviour related to heavy breathing. On Day 0, they were allocated into one of two treatments, viz, Infected, in which animals were orally dosed weekly for three weeks with a 1:1 mixture of 20,000 O. ostertagi and C. oncophora L3 larvae, or Control in which animals remained uninfected, and grazed a new pasture that had not previously been grazed. Animals were rotationally grazed as a mob with weekly shifts and each break back-fenced to reduce the risk of self-infection. On Day 49, to assess the behavioural response following anthelmintic treatment, half of calves in each treatment were orally administered with the Matrix Oral triple combination anthelmintic and continued to be monitored, using a 2 × 2 factorial experimental design, viz, Control without Drench (COD), Control with Drench (CD), Infected without Drench (IOD), Infected with Drench (ID). Results showed nematode challenge increased daily eating time (+14.8 minutes per day, mins/d, P = 0.007) and reduced daily ruminating time (-9.8 mins/d, P = 0.001) and daily mid-activity duration (-5.3 mins/d, P = 0.023) in grazing calves. After 49-day challenge of gastrointestinal nematode (GIN) infection, the mean faecal egg counts (FEC) of Infected and Control groups were 700 ± 181 and 33 ± 18 eggs/gram, respectively. Furthermore, there was a lack of a consistently high level of parasitism in the calves. To help determine if there was any potential for subtle changes in behaviour to be detected, three animals at the extremes of each of the initial intended treated groups, viz, animals in IOD and 3 animals in ID with the highest FEC and three calves in COD and four calves in CD with highest body weight gain were selected. Parasitism increased FEC and decreased daily live weight gain (643.7 g/day vs 790.6 g/day for parasitized animals and control animals respectively, P = 0.029) from Day 20 to Day 49. There were non-significant negative correlations between log10 (FEC+1) and behavioural measurements for daily eating duration (r = -0.085, P = 0.501), daily walking duration (r = -0.16, P = 0.191), daily mid-activity duration (r = -0.18, P = 0.155), daily high activity duration (r = -0.060, P = 0.636), and daily heavy breathing duration (r = -0.17, P = 0.174) with non-significant positive correlations for daily ruminating duration (r = 0.018, P = 0.889) and daily rest duration (r = 0.16, P = 0.197). In the parasitized animals with the high-level output of FEC, nematode infection significantly increased their mean daily eating duration (+34.6 mins/d, P = 0.020), compared with their uninfected counterparts, and had no effect on mean daily ruminating duration (-18.2 mins/d, P = 0.737), mean daily walking duration (-14.4 mins/d, P = 0.598), mean daily mid-activity duration (-10.3 mins/d, P = 0.498), mean daily high activity duration (-7.2 mins/d, P = 0.744). Moreover, nematode infection increased eating time during cooler periods of the day (+1.7 mins/h from 0400 to 0800 hours, +1.5 mins/h from 2000 to 2300 hours) and decreased in the early morning (-0.7 mins/h from 0000 to 0300 hours) and in the afternoon (-1.5 mins/h from 1300 to 1700 hours) (P < 0.001), and reduced the duration of mid-activity during most times of the day (-0.5 mins/h from 0000 to 0700 hours, -0.2 mins/h from 1100 to 1300 hours, -0.2 mins/h from 1500 to 2300 hours, and +0.2 mins/h from 0800 to 1000 hours, P = 0.007). In parasitized animals post anthelmintic drenching, heavy breathing duration was decreased (-0.1 mins/h) from 0200 to 0500 hours and then increased (P = 0.010) during most time of the day (from 0600 to 0900 hours, from 1100 to 1600 hours and from 2200 to 0100 hours) with fluctuations from 1700 to 2100 hours. In control animals post anthelmintic drenching, walking duration was reduced (P = 0.010) during early morning from 0100 to 0700 hours (-0.4 mins/h) and at dusk from 1700 to 1800 hours (-0.6 mins/h), and increased from 0800 to 1600 hours (+0.3 mins/h) and in the late evening from 1900 to 0000 hours (-0.2 mins/h), while heavy breathing duration was increased (P = 0.018) during most times of the day (+0.1 mins/h from 0500 to 0900 hours, +0.2 mins/h from 1100 to 0000 hours). Unfortunately due to an outbreak on Mycoplasma bovis on the research farm the first study was required to be terminated after only a short-time post-drenching. Due to biosecurity requirements continuing investigations in cattle was not an option and it was decided to instead use sheep as the species of choice for the remainder of the studies. A paucity of commercially available sensors for sheep meant that alternatives must be found and validated. Of these, the Cowmanager sensor tags were chosen due to their relative size and weight and possible suitability for a sheep ear. However, these tags had not previously been evaluated in sheep. The second trial (Chapter 4) was aimed to determine if the Cowmanager tags could provide an accurate and meaningful representation of the behaviour of sheep. In order to do this, a lipopolysaccharide (LPS) challenge model was used which provides a predictable and consistent short-term change in animal behaviour such as lethargy, and inappetence without long-term negative consequences. At recruitment (Day -10), twenty female Coopworth lambs of 8-10 months (mean live weight = 38.63 ± 2.04 kg) were weighed and fitted with CowManager SensOor ear tags which can identify mutually exclusive activities of five categories: eating, ruminating, inactive behaviour, active behaviour and highly active behaviour, and allocated within live weight strata to one of two treatments, respectively receiving 0 or 0.5 μg/kg body weight of endotoxin LPS with an experimental design of 2 × 2 Latin square. On Day 0, the lambs in the LPS group were intravenously injected with LPS contained in sterile saline, while the control lambs were intravenously injected with sterile saline solution. Seven lambs (four lambs in the LPS group and three control lambs) were randomly selected and individually marked for visual identification for verification of the behavioural changes. To validate the data between visual observation and sensor recordings, the behaviours of the seven lambs were visually observed and recorded every minute for 60 consecutive minutes for two hours between 0900 hours and 1600 hours within three days around the period of LPS infusion. A week after the first intravenous infusion of LPS, the two groups were swapped and intravenously administered with 0.5 or 0 μg/kg body weight of LPS, and the experimental procedure was repeated for the second period. The behaviours of the same seven lambs previously marked for visual identification was recorded via visual observation based on the same procedure to compare with sensor recordings. LPS infusion elevated rectal temperatures after 4 hours from 39.31℃ to 39.95℃, indicating successful establishment of an acute fever response for comparison between groups (P < 0.001). Results showed there was good agreement between visual observations and sensors for active and not active behaviour, but poor agreement with eating and ruminating time. For each of the five recorded behaviours, time spent eating, ruminating, not active, active and highly active, the accelerometers were able to detect an effect of LPS challenge. Compared with the control, LPS infusion decreased eating time (-6.7 mins/h, P < 0.001), active behaviour (-8.4 mins/h, P < 0.001) and highly active (-2.9 mins/h, P < 0.001) and rumination time (-1.4 mins/h, P = 0.075) and increased inactive behaviour (+16.0 mins/h, P < 0.001) in challenged lambs. This provided validation of the Cowmanager tags for use in sheep. Alkaloids produced by ryegrass endoyphytes are known to cause changes in animal behaviour, inducing heat stress and ryegrass staggers in severe cases. In Chapter 5, the potential for CowManager SensOor ear tags to categorize changes in eating, ruminating and other behavioural activities of grazing lambs exposed to endophyte-infected perennial ryegrass was evaluated. Thirty Coopworth lambs with a mean live weight (± SE) of 33.6 ± 0.46 kg were allocated randomly within live weight strata to either endophyte-free cultivars of ryegrass (Control) or wild type endophyte-infected cultivars of ryegrass pasture (Endophyte). Live weight change, behaviour and incidence of ryegrass staggers were monitored over a 2-month grazing period. Moderately severe staggers (score 4/5) occurred in 40% of lambs in the endophyte treatment with a mean staggers score of 2.33 ± 0.41 (P < 0.001) for the endophyte group at the end of trial. During the period of ryegrass staggers, compared with those grazing the control pasture, lambs in the endophyte group had had no significant difference in eating time (-36.0 mins/d, P = 0.516), time spent being active (+29.9 mins/day, P = 0.556) and being inactive duration (-43.7 mins/day, P = 0.114), but increased time spent ruminating (+40.2 mins/d, P = 0.051). When comparing lambs in the same group with stagger score 4 and those with staggers score 1 or 0 during staggers period, these behavioural differences were even more pronounced though non-significance was presented (-82.2 mins/d for eating, P = 0.073; -62.9 mins/d for being inactive, P = 0.221; +124.0 mins/d for being active, P < 0.001; +48.9 mins/d for ruminating, P = 0.015), indicating that it was not a reflection of the different swards or the physical characteristics. Changes in the pattern of behaviours during staggers period were also evident as reflected in a diurnal alteration of each activity over time of the day. The challenge of toxic alkaloids produced in endophyte-infected ryegrass led to more time eating during cooler period of the day (P < 0.001) and more time being active for compensation during day-time (P < 0.001) as well as more ruminating time during night-time (P < 0.001) and less time being inactive (P < 0.001) and highly active (P < 0.001), especially in the lambs with severe ryegrass staggers. The final trial then compared the effect of a chronic subclinical infection with gastrointestinal nematode parasites on sheep behaviours (Chapter 6). At start of this trial on Day 0, thirty-six Coopworth ram lambs at 8-10 months of age were weighed, fitted with CowManager SensOor ear tags, and randomly allocated within live weight to one of four experimental groups, non-parasite exposure groups (NP), NP1 (n = 9, 34.89 ± 3.93 kg,) and NP2 (n = 9, 34.56 ± 2.82 kg), oral administration of parasites (OP), OP1 (n = 9, 34.89 ± 3.00 kg) and OP2 (n = 9, 34.44 ± 2.94 kg) with a three-times weekly trickle infection for four weeks with the equivalent of 130 Teladorsagia (Ostertagia) circumcincta L3 larvae per kg live weight (LW) and 80 Trichostrongylus colubriformis L3 larvae per kg LW per day. After 4 weeks of infection, the NP1 group and OP1 group were orally administered once with an anthelmintic. Two weeks later, the NP2 group and OP2 group were orally treated once with the same anthelmintic. GIN infection resulted in a reduction in mean daily eating time (-91.7 mins/d, P = 0.001) and an increase in mean time spent being inactive (+88.4 mins/d, P = 0.001) and active (+7.4 mins/d, P = 0.017) with no differences in mean daily ruminating time (-15.5 mins/d, P = 0.395) and mean daily highly active duration (+12.7 mins/d, P = 0.400). Compared with the control, the animals challenged with nematode infection increased eating time during cooler period of the day (+1.5 mins/h from 2000 to 0200 hours) with less time spent eating during sunrise (-4.7 mins/h) and sunset (-1.7 mins/h) (P < 0.001). Further, compared with the control, the animals infected with nematode reduced daytime ruminating (-0.5 mins/h from 0600 to 1600 hours) and increased ruminating time at dusk (+0.5 mins/h from 1700 to 1900 hours) and in the evening (+0.8 mins/h from 2100 to 0000 hours) (P < 0.001). In addition, compared with the control, nematode infection increased inactive time of parasitized animals in the early morning/after sunset with a reduction during the late afternoon (-0.6 mins/h from 1400 to 1800 hours) and late evening (-1.1 mins/h from 2200 to 0100 hours) (P < 0.001), increased active duration in the early morning (+1.2 mins/h from 0400 to 0700 hours) with reductions in the late morning (-0.2 mins/h from 1000 to 1200 hours) and from late afternoon to evening (-0.3 mins/h from 1500 to 2200 hours) (P < 0.001), and increased highly active duration during most times of the day (+0.6 mins/h from 0300 to 1800 hours) with a decrease in the evening (-1.1 mins/h from 1900 to 0200 hours) (P < 0.001). Compared with the control animals, for the infected animals, anthelmintic drenching increased eating time during most of the daytime +1.6 mins/h with a decrease during cooler times of the day (-2.4 mins/h from 2000 to 0300 hours) (P < 0.001), increased night-time ruminating (+1.3 min/h from 2000 to 0200 hours) with an increase in the early morning (-1.1 mins/h from 0300 to 0800 hours) and in the afternoon (-1.0 mins/h from 1300 to 1800 hours) (P < 0.001), increased inactive duration in the morning (+0.9 mins/h from 0200 to 1000 hours) and in the afternoon (+1.3 mins/h from 1200 to 1700 hours) with a decrease in the early evening (P < 0.001), reduced active duration from 0300 to 0500 hours (-0.6 mins/h) and from 0700 to 1400 hours (-1.1 mins/h) with an increase from 1500 to 1800 hours (+0.9 mins/h) and from 2000 to 0200 hours (+0.9 mins/h) (P < 0.001), decreased highly active duration during most of the daytime (-1.1 mins/h from 1000 to 1800 hours) with an increase in the evening (+2.2 mins/h from 1900 to 0100 hours) (P < 0.001). In summary, the results from these trials demonstrate that altered behavioural patterns of grazing animals due to health challenges of diseases or toxins can be detected through commercially available sensor technologies. Acute health challenge in lambs can increase inactive duration and compensate to reduce eating duration, rumination duration, active duration and highly active duration in a short period, whereas chronic endophyte staggers challenge in lambs results in increased daily active duration and daily ruminating duration with a decrease in daily eating duration, daily inactive duration and daily highly active duration. Chronic subclinical parasitic challenge in dairy calves can induce an increase in daily eating duration and daily rest duration, and decrease daily durations of ruminating, walking, mid-activity, high-activity and heavy breathing. However, chronic subclinical nematode challenge in lambs can result in a decrease in daily durations of eating, ruminating and an increase in the daily durations of being inactive, active and highly active. The magnitude and duration of total daily behavioural changes varied under these health challenges. Moreover, both chronic subclinical GIN challenge and chronic challenge of toxic alkaloids in endophyte-infected ryegrass can increase eating time during cool periods of the day, decrease time spent being active during nighttime with an increase in the daytime in grazing lambs at pasture. However, chronic subclinical GIN challenge can increase inactive duration at sunrise and sunset and decrease daytime ruminating and increase ruminating duration in the evening, while chronic challenge of toxic alkaloids can increase ruminating duration with a reduction in inactive duration during most of the day. Consequently, behavioural changes detected by commercially available sensor technologies evaluated in these studies could be applied as a useful and potential tool to identify changes in behavioural patterns of each category under infection challenges of diseases or toxins, indicating animal health status and determining the need of treatments for individuals or a whole mob.