Bruna F. Silper

Short communication: Factors affecting hair cortisol concentrations in lactating dairy cows.

Cortisol has long been used as a marker of the stress response in animals. Cortisol can be analyzed from different media, most notably from the blood, saliva, and feces; however, the collection of cortisol from some of these media requires invasive procedures or excessive handling of the animals. Furthermore, it is not possible to capture long-term increases in circulating concentrations of cortisol from the blood, saliva, or feces. Hair cortisol has been found to be a reliable alternative for measuring chronic stress. With this emerging measure, appropriate sampling methodology must be developed and validated. The aim of this study was to determine the effects of hair color, sampling location, and processing method on cortisol concentrations in hair from lactating black and white Holstein cows (n=18). Furthermore, we aimed to measure the hair growth rates at different body locations (n=12) and test hair cortisol levels when resampled over short intervals (n=37). Both black- and white-colored hair was collected from the shoulder, top line, hip, and tail switch of Holsteins; due to breed characteristics only white hair was harvested from the tail switch. All samples were cleaned with water and isopropanol, and then ground in a ball mill or finely cut with scissors once dry. Cortisol was extracted with methanol before being measured using a commercially available ELISA kit. Concentrations of cortisol were greater in white than in black hair (7.8 ± 1.1 vs. 3.8 ± 1.1 pg/mg). When only white samples were analyzed, hair from the tail switch had more cortisol than hair from the shoulder (11.0 ± 1.2 vs. 6.2 ± 1.2 pg/mg), whereas no difference was found when compared with the hip and top line. Samples ground with a ball mill had greater concentrations of cortisol extracted than those minced with scissors (10.4 ± 1.2 vs. 4.7 ± 1.2 pg/mg). The growth rate of hair was significantly greater at the tail switch compared with the hip and shoulder (0.51 ± 0.05 vs. 0.04 ± 0.05 vs. 0.03 ± 0.05 mm/d). When hair was collected every 3 wk after calving, a tendency was detected for multiparous cows to have greater concentrations of hair cortisol and significantly greater concentrations of cortisol on d 0 and 21 after calving compared with d 42, 84, and 126. In Holsteins, the hair on the tail switch is always white, grows more rapidly than other sites, and is sensitive enough to capture changes in cortisol over intervals as short as 3 wk, making it the ideal location for measuring hair cortisol.

Factors affecting expression of estrus measured by activity monitors and conception risk of lactating dairy cows

The objective of this study was to determine risk-factors affecting increase in physical activity during estrus and pregnancy per artificial insemination (P/AI) in lactating dairy cows. Cows were monitored continuously by 2 automated activity monitors [a collar-mounted accelerometer (HT; Heatime, SCR Engineers, Netanya, Israel) and a leg-mounted pedometer (BO; Boumatic Heat-seeker-TX, Boumatic Dairy Equipment, Madison, WI)]. When an increase in activity was detected, body condition score (BCS) and blood samples were collected, ovaries were scanned by ultrasonography, and, if the cow was eligible for breeding, artificial insemination was performed. Milk production and health-related data were recorded throughout the experimental period. Pregnancy diagnosis was performed at 42 ± 7 d of gestation. Data were analyzed using Pearson correlation, ANOVA, and logistic regression. A total of 1,099 true events of estrus from 318 lactating Holstein cows were recorded, averaging 3.46 ± 1.1 events per cow. Positive predictive value for estrus episodes detected by the HT and BO systems were 89.6 and 85.5%, respectively. Mean peak activity at estrus (PA) recorded by the HT system was 71.6 ± 20.7 index-value, and 334.3 ± 155.7% relative increase by the BO system. Compared with primiparous, multiparous cows expressed estrus with lower PA (69.3 ± 0.8 vs. 75.9 ± 1.1 index for HT; 323.9 ± 6.0 vs. 354.8 ± 8.48% for BO) and shorter duration (DU; 10.7 ± 0.2 vs. 12.0 ± 0.3 h); DU was measured by HT only. Lower BCS was associated with decreased PA measured by both systems, estrus DU, and P/AI. Peak activity was weakly correlated with milk production on the day of artificial insemination (r = -0.20); however, when categorized into quartiles, the highest-yield cows had lower PA and DU. Follicle diameter was not correlated with PA or DU, but cows with greater concentrations of estradiol had higher PA. Cows with greater PA in both systems had greater P/AI than those with lower PA (36.5 vs. 24.6% for HT, 33.5 vs. 21.4% for BO). In conclusion, measurements of estrus events captured by automated activity monitors are correlated with BCS, parity, and secondary behavior signs related to estrus. Surprisingly, estrus intensity and duration were only weakly correlated with milk production, preovulatory follicle diameter, and concentrations of estradiol at estrus. Cows that had measurements of high-intensity estrus were significantly more fertile than low-intensity estrus.

Automated and visual measurements of estrous behavior and their sources of variation in Holstein heifers. I: Walking activity and behavior frequency

Holstein heifers (n = 57) were monitored using accelerometers and video observations with the objective of better understanding the behavioral expression of estrus, the variation within and between the heifers, and the possible sources of variation. IceTags recorded walking activity from 7 to 13 months of age. Activity peaks (n = 282) were obtained from a rolling sum of steps within 24-hour periods and validated to be estrus by ovarian ultrasonography. Behavior around activity peak of one estrus for each of 12 heifers was described in detail from video recordings. Baseline behavior was monitored in a corresponding interval 1 week before. Estrus and baseline total steps and steps per hour, estrus relative increase in activity, duration, and interval between episodes were analyzed by descriptive statistics and Spearman rank correlations. Effects of category of baseline walking activity, estrus order (pubertal vs. second and greater episodes), season, hour of estrus onset, and number of heifers simultaneously in estrus were evaluated with proc MIXED. Behavioral changes from baseline to estrus were evaluated by a signed-rank test. Estrus total steps varied greatly (4743 ± 1740; range: 837-10,070), as well as the relative increase in activity (290 ± 160%; range: 30%-1190%). Duration of estrus was 14 ± 4 hours, ranging from 4 to 26 hours. The interval between episodes was the trait that varied the least. Pubertal estrus was shorter and had a smaller relative increase in activity than second and greater episodes (P < 0.05). The number of steps during estrus was greater for heifers of high baseline activity (P < 0.01). Estrus episodes occurring in the winter and starting between 4 PM and 3 AM had the greatest relative increase in activity (P < 0.05). The number of heifers simultaneously in estrus did not influence estrus expression (P > 0.05). The behaviors with greatest change from baseline to estrus were chin rest, sniff, back mount, crossover, accept chin rest, and follow, but variation was large. Overall, estrus was apparent in behavioral changes with large variation within and between the heifers. Estrus order, onset hour, season, and baseline walking activity are important factors affecting estrus activity. Therefore, estrus detection tools should account for potential sources of variation. The visual and automated measurements of estrus expression reported in this study reveal possibilities for improved on-farm estrus detection technologies and potential genetic selection for estrus expression.

Relationship of concentrations of cortisol in hair with health, biomarkers in blood, and reproductive status in dairy cows

Hair cortisol has been used to measure chronic stress in dairy cows as it offers the advantage of being noninvasive, fast, and able to indicate levels of cortisol over long periods. The aim of this study was to determine the associations between hair cortisol with clinical disorders, reproductive status, and the development of subclinical endometritis in dairy cows. Furthermore, we aimed to determine the association between hair cortisol concentrations and blood markers associated with metabolic status and acute inflammation. In experiment 1, cows (n=64) were hair sampled every 3wk from the tail switch beginning at calving (d 0) until d 126 for cortisol analysis; blood samples were collected every 3wk from d 0 until 42 for β-hydroxybutyrate and glucose analysis. In experiment 2, cows (n=54) were chosen retrospectively by diagnosis of subclinical endometritis (END), subclinical endometritis and at least 1 clinical disease (END+CLIN), or as healthy (control) using a cytobrush and ultrasonography at 30±3d in milk. At the same time, animals were hair sampled for cortisol analysis and blood sampled for haptoglobin and ceruloplasmin analysis. Health records were recorded throughout both experimental periods. Animals with clinical disease presented higher cortisol concentrations than clinically healthy animals in experiment 1 [geometric mean (95% confidence interval); 8.8 (7.8, 9.9) vs. 10.7 (9.6, 12.0) pg/mg]; however, animals diagnosed with subclinical endometritis in experiment 2 did not differ in hair cortisol concentrations [11.7 (9.8, 14.0), 12.2 (9.3, 15.9), 10.5 (8.1, 13.6) pg/mg for control, END, and END+CLIN, respectively]. In experiment 1, an effect of sample day was noted, where d 21 had higher cortisol concentrations than d 42, 84, and 126, but not from d 0 for both parities. Within both experiments, a parity effect was present where multiparous animals consistently had higher cortisol concentrations than primiparous animals. Multiparous cows that became pregnant by 100d postpartum had lower concentrations of hair cortisol at d 42 and 84 in milk. Lastly, other biomarkers associated with metabolic status and acute inflammation, such as glucose, β-hydroxybutyrate, haptoglobin, and ceruloplasmin, were not strongly correlated with measurements of cortisol in hair. Overall, hair cortisol measurements appear to be associated with clinical disorders and have a direct association with pregnancy status; however, concentrations of hair cortisol may not be suited to differentiate situations of stress with lower magnitudes, such as the development of subclinical disease.

Short communication: Comparison of estrus characteristics in Holstein heifers by 2 activity monitoring systems

Two activity monitoring systems-Heatime (SCR Engineers Ltd., Netanya, Israel) and IceTag (IceRobotics Ltd., Edinburgh, UK)-were compared on their ability to detect and quantify estrus expression. Holstein heifers (n=57) were fitted with Heatime (HT) and IceTag (IT) sensors from 12 mo of age until confirmation of pregnancy. Upon detection of high activity by HT, ovaries were scanned by ultrasound, a blood sample was collected for analysis of plasma estradiol, and signs of estrus (clear vaginal mucus, uterine muscle tone, visual mounting activity, standing to be mounted, or rump showing signs of repeated acceptance of mounts) were recorded. Because only estrus episodes detected by HT (n=111) were further evaluated, only the positive predictive value was measured. Heifers were housed in groups of 24 in a freestall pen. Data were analyzed using Proc CORR and GLM of SAS (SAS Institute Inc., Cary, NC). The positive predictive value was 84.7% (94/111) for HT and 98.7% (74/75) for IT. Estrus duration was recorded by HT as 14.3±4.1h [mean ± standard deviation (SD)] and by IT as 15.0±4.0h; duration measurements were correlated (r=0.60). The mean duration difference was 0.74±3.52h. Recordings of onset and end of estrus by IT were 3.5±4.3h and 2.9±4.9h earlier than those by HT. The overlap in duration was 9h. Measurements of estrus intensity were correlated (r=0.63). Peak activity was 77.3±19.5 index value (approximately 7.7 SD from basal activity) on HT. The relative increase in activity measured by IT was 360±170% baseline value. Measurements of intensity and duration from HT were correlated (r=0.64) but those from IT were not (r=0.13). Plasma estradiol concentration (11.2±4.6pg/mL) was not correlated with preovulatory follicle diameter or with duration or intensity of estrus. Diameter of preovulatory follicle (15.7±2.6mm) had no correlation with duration of estrus and was only weakly correlated with intensity measured by either system. Baseline steps/hour was negatively correlated with intensity from both sensors (r=-0.37 and -0.70 for HT and IT). Estrus episodes accompanied by 2 or 3 of the monitored signs of estrus had greater intensity and duration on HT but not on IT. Preovulatory follicle diameter and plasma estradiol concentration did not influence occurrence of estrus signs. Results indicate that both systems identified estrus precisely, with correlated characterization and similar timing. In contrast, relationships with plasma estradiol concentration and signs of estrus require further investigation.

Automated and visual measurements of estrous behavior and their sources of variation in Holstein heifers. II: Standing and lying patterns

Lying (LY) and standing (ST) behavior of Holstein heifers (n = 57) was studied from Day -7 to Day +2 relative to estrus (Day 0) using leg-mounted accelerometers. A total of 269 estrus episodes were studied. The objectives were to quantify the effects of estrus on ST and LY patterns, identify sources of variation, and provide insight into new approaches for the use of sensors in estrus detection. The sensors were used to obtain ST bout frequency (bouts per day), total daily ST time, and mean bout duration and duration of the longest bout for ST and LY on a daily basis. Estrus episodes were identified using walking activity peaks and validated by ovarian ultrasonography. Repeated-measures ANOVA was used to analyze the effects of estrus order (pubertal vs. second and greater), season, time of estrus onset, category of baseline walking activity, and number of heifers simultaneously in estrus on ST and LY variables. Overall, ST bout frequency was lower, whereas mean ST bout duration, duration of the longest ST bout, and total daily ST time were longer on Day 0 (P < 0.05). The longest ST bout increased from 232 ± 5 minutes on Day -7 to 488 ± 16 minutes on Day 0 (P < 0.05), and on Day 0, it started within -2 to 4 hours of time of estrus onset. Measurements of ST bouts varied the most on Day 0 (coefficient of variation = 21.5%-52.8%). The pubertal estrus was characterized by smaller increase in mean ST bout duration, longest ST bout duration, and total daily ST time (P < 0.05) compared to the second and greater episodes. Second and greater episodes had greater longest LY bout on Day +1, but on Day 0 it was not different from other nonestrus days (P < 0.01). The longest ST bout on Day 0 was longer for heifers with high baseline walking activity and for episodes occurring during the cold season (P < 0.05). The effect of estrus on ST variables for morning-onset episodes was observed only on Day 0. Episodes with afternoon and night onset showed the effect of estrus spread over Days -1 and 0, resulting in a smaller change between Day -7 and Day 0. The number of heifers simultaneously in estrus did not affect the studied variables (P > 0.05). Measurements of ST and LY changed during estrus and were correlated to walking activity. The large increase in duration of the longest ST bout and its occurrence in proximity of estrus onset suggest potential for automation. Incorporating measurements of ST and LY corrected for sources of variation could improve estrus detection systems.

Integrating an automated activity monitor into an artificial insemination program and the associated risk factors affecting reproductive performance of dairy cows

The aim of this study was to compare 2 reproductive programs for the management of first postpartum artificial insemination (AI) based on activity monitors and timed AI, as well as to determine the effect of health-related factors on detection and expression of estrus. Lactating Holstein cows (n = 918) from 2 commercial farms were enrolled. Estrous cycles of all cows were presynchronized with 2 injections of PGF2α administered 2 wk apart. Treatments were (1) first insemination performed by timed AI (TAI) and (2) first insemination based upon the detection of estrus by activity monitors (ACT; Heatime, SCR Engineering, Netanya, Israel) after the presynchronization, whereas cows not inseminated by the detection of estrus were enrolled in the Ovsynch protocol. Body condition score (BCS; scale 1 to 5), hock score (scale: 1 to 4), gait score (scale: 1 to 4), and corpus luteum presence detected by ovarian ultrasonography were recorded twice during the presynchronization. On the ACT treatment, 50.5% of cows were inseminated based on detected estrus, whereas 83.2% of the cows on the TAI treatment were inseminated appropriately after the timed AI protocol. Pregnancy per AI did not differ by treatment (30.8 vs. 33.5% for ACT and TAI, respectively). Success of pregnancy was affected by parity, cyclicity, BCS, milk production, and a tendency for leg health. In addition, treatment × cyclicity and treatment × parity interactions were found to affect pregnancy success, where anovulatory cows and older cows had compromised pregnancy outcomes on the ACT treatment but not on the TAI treatment. Factors affecting pregnancy outcomes varied among farms. Hazard of pregnancy by 300 DIM was affected by farm, parity, BCS, a treatment × cyclicity interaction, and a tendency for an interaction between leg health and farm. Detection of estrus was affected by farm, parity, cyclicity, and leg health, but not BCS or milk production. Expression of estrus was compromised in anovular and older cows, and by the timing of the estrus event, but not by gait score, BCS, or milk production. Increased duration of estrus, but not intensity of estrus, improved pregnancy per AI. In conclusion, using an automated activity monitor for the detection of estrus within a Presynch-Ovsynch program resulted in similar pregnancy per AI and days open compared with a reproduction program that was strictly based on timed AI for first postpartum AI. In contrast, notable variations in reproductive outcomes were detected between farms, suggesting that the use of automated activity monitors is prone to individual farm management.

Daily lying behavior of lactating Holstein cows during an estrus synchronization protocol and its associations with fertility

Objectives were to quantify lying behavior (LB) during an estradiol and progesterone-based synchronization protocol, to assess risk factors for ovulation, pregnancy per AI (P/AI), and degree of behavioral change at estrus, and to investigate the associations between estrus LB and walking activity. Holstein cows (43.6 ± 11.0 kg of milk/d) were fitted with leg-mounted accelerometers. Total lying time/d (L_time), bout frequency (bout_N), average lying bout duration, and relative increase in walking activity (ACT%) were evaluated for 1,411 timed artificial insemination events. The day with lowest L_time or bout_N among d -2, -1, and 0 (day of timed artificial insemination) determined the day of behavioral estrus. The variables L_time% and bout_N% represent relative ratios between lowest L_time and baseline (d -7), L_time, and lowest bout_N, and baseline (d -7) bout_N, respectively [e.g., (lowest L_time/baseline L_time) × 100]. Correlation coefficients between L_time% and bout_N% and ACT% were -0.38 and -0.31, respectively. Estrus LB change was considered large if <75% of baseline and small if ≥75% of baseline for both L_time% and bout_N%; average lying bout duration did not change with estrus. Lowest L_time% and bout_N% corresponded to, respectively, 65 ± 21% (mean ± standard deviation; 447 ± 157 min/d) and 65 ± 24% (8.5 ± 4.0 bouts/d) of baseline. The change in L_time% at estrus was smaller when cows had milk yield above average; the change in bout_N% was smaller among multiparous cows and for estrus occurring in the colder season. Likelihood of ovulation was greater when there was larger change in L_time% [odds ratio = 4.9; ovulation rate = 93 (large change) and 76% (small change)], as well as when a corpus luteum was present at start of protocol (odds ratio = 3.6; in the model with L_time%). Likelihood of pregnancy at d 32 was 1.6 times greater for estrus with large change in LB [L_time% or bout_N%; P/AI = 34% (large change in L_time%) and 26% (small change in L_time%)]. Among estrus events with ACT% ≥300% (high intensity), classification by small or large L_time% did not influence P/AI at 32 d. The magnitude of LB change at estrus and its association with fertility suggest potential application toward improved use of activity monitors (e.g., increased estrus detection, fertility prediction). The contribution of LB to accuracy of estrus detection when physical activity is known remains to be addressed. The relationship between intensity of estrus expression and fertility requires further investigations of its physiological rationale and on-farm applications.

Diagnosis of uterine and vaginal disorders by different methodologies is affected by concentration of estradiol in plasma from lactating Holstein cows.

The relationship between plasma estradiol concentration at time of examination and prevalence of uterine disorders, agreement among methods, and associations of diagnosis with pregnancy hazard and milk yield was studied in 268 Holstein cows examined at 30±3 (exam 1) and 44±3 d in milk (DIM; exam 2). Purulent vaginal discharge was sampled using 2 methods: gloved hand and Metricheck (Simcro, Hamilton, New Zealand; PVD; score ≥3). Percentage of polymorphonuclear leukocytes was determined by endometrial cytology (CYTO; exam 1: ≥18%, exam 2: ≥10%); diameter of uterine horns (UTH; >20 mm), diameter of the inner layer of the cervix (CVX; >20.5 mm), presence of fluid in the uterine lumen (FL), and ovarian structures were evaluated by ultrasonography. A blood sample was collected at each exam for estradiol analysis. Prevalence at exams 1 and 2 was, respectively, 14.2 and 18.5% (PVD), 21.4 and 10.1% (FL), and 40.6 and 50.2% (CYTO). Prevalence of PVD at exam 1 was greater among cows with estradiol ≥2 pg/mL (19.4 vs. 8.2%). Agreement of all methods with CYTO was poor, the greatest being between CYTO and FL (exam 1; kappa=0.19). Agreement between CYTO and PVD, and between CYTO and FL (exam 1; kappa=0.15 and 0.35, respectively) was higher among cows with estradiol ≥2 pg/mL. Likelihood of PVD at exam 1 was greater if cows were positive for CVX [odds ratio (OR)=3.0], FL (OR=2.6) or had estradiol ≥2 pg/mL (OR=2.7). Likelihood of CYTO increased with dystocia (OR=2.3) and FL (OR=2.5). Estradiol did not influence diagnosis at exam 2. Positive FL or CYTO at exam 1 was associated with reductions in milk yield of 59 to 180 kg by 45 DIM. Pregnancy hazard until 250 DIM was reduced by CYTO at exam 1 (hazard ratio=0.74) and by PVD (hazard ratio=0.68) at exam 2. However, FL and CYTO reduced pregnancy hazard only when estradiol was ≥2 pg/mL (exam 1), whereas PVD reduced pregnancy hazard when diagnosed at exam 2 with estradiol <2 pg/mL. Overall, agreement was poor and effects of positive diagnosis differed according to method and DIM at exam. Estradiol concentration influenced prevalence, agreement, likelihood of positive diagnosis, and its effects on days to pregnancy.

Pilot Study to Evaluate the Association Between the Length of the Luteal Phase and Estrous Activity Detected by Automated Activity Monitoring in Dairy Cows

The ability of automated activity monitoring (AAM) systems to detect estrus is influenced by multiple variables. The luteal phase (LP) preceding estrus has been shown to be important for hormones release, and behavior during estrus in ruminants, but its impact on activity as measured by AAM systems has not been studied. The objective of this study was to investigate the impact of the length of the preceding LP on the intensity and duration of estrus as measured by AAM, and on the expression of estrogen receptor alpha (ERα) in the endometrium. A total of 60 cows between 46 and 53 days in milk were assigned to either a normal or a short length LP. Length of the LP was insured by the use of hormonal synchronization programs. Cows in both treatments were matched on parity, milk production, body condition score (BCS), and lameness score (assessed at enrolment). Expression of ERα receptors in the endometrium were evaluated by nuclear staining by immunohistochemistry of biopsies. Estrus was defined as the day on which the AAM system detected estrus. Cows that were not synchronized by the protocol or that were not detected in estrus by the AAM systems were excluded, which left 21 and 11 cows for analyses in the normal and short length LP, respectively. Peak activity index of estrus, duration of estrus, and expression of ERα were outcomes for multivariable linear regression models. Cows with short length LP tended to have lower peak activity at estrus, but there was no significant effect of treatment when BCS was accounted for. Cows with BCS ≤ 2.5 had less intense and shorter estrus than cows with BCS ≥ 2.75. There was no association between the length of the LP and the expression of ERα in the endometrium.