- Email: [email protected]
Effect of heat stress on body temperature in healthy early postpartum dairy cows
Available online at www.sciencedirect.com
Theriogenology 78 (2012) 2031–2038 www.theriojournal.com
Effect of heat stress on body temperature in healthy early postpartum dairy cows O. Burfeind, V.S. Suthar, W. Heuwieser* Clinic of Animal Reproduction, Faculty of Veterinary Medicine, Freie Universität Berlin, Berlin, Germany
Abstract Measurement of body temperature is the most common method for an early diagnosis of sick cows in fresh cow protocols currently used on dairy farms. Thresholds for fever range from 39.4 °C to 39.7 °C. Several studies attempted to describe normal temperature ranges for healthy dairy cows in the early puerperium. However, the definition of a healthy cow is variable within these studies. It is challenging to determine normal temperature ranges for healthy cows because body temperature is usually included in the definition. Therefore, the objectives of this study were to identify factors that influence body temperature in healthy dairy cows early postpartum and to determine normal temperature ranges for healthy cows that calved in a moderate (temperature humidity index: 59.8 ⫾ 3.8) and a hot period (temperature humidity index: 74.1 ⫾ 4.4), respectively, excluding body temperature from the definition of the health status. Furthermore, the prevalence of fever was calculated for both periods separately. A subset of 17 (moderate period) and 15 cows (hot period) were used for analysis. To ensure their uterine health only cows with a serum haptoglobin concentration ⱕ 1.1 g/L were included in the analysis. Therefore, body temperature could be excluded from the definition. A vaginal temperature logger that measured vaginal temperature every 10 min was inserted from Day 2 to 10 after parturition. Additionally rectal temperature was measured twice daily. Day in milk (2 to 10), period (moderate and hot), and time of day had an effect on rectal and vaginal temperature. The prevalence of fever (ⱖ 39.5 °C) was 7.4% and 28.1% for rectal temperature in the moderate and hot period, respectively. For vaginal temperature (07.00 to 11.00 h) it was 10% and 33%, respectively, considering the same threshold and period. This study demonstrates that body temperature in the early puerperium is influenced by several factors (day in milk, climate, time of day). Therefore, these factors have to be considered when interpreting body temperature measures to identify sick cows. Furthermore, the prevalence of fever considering different thresholds is higher during hot than moderate periods. However, even in a moderate period healthy cows can exhibit a body temperature that is considered as fever. This fact clearly illustrates that fever alone should not be considered the decision criterion whether a cow is allocated to an antibiotic treatment, although it is the most important one that is objectively measurable. © 2012 Elsevier Inc. All rights reserved. Keywords: Dairy cow; Body temperature; Fever; Climate; Haptoglobin; Fresh cow protocol
1. Introduction Currently, measurement of body temperature (BT) is one of the most common methods used for the diagnosis of infectious diseases (i.e., puerperal metritis, clinical mastitis) in the early puerperium in dairy cows [1]. * Corresponding author. Tel.: ⫹49 30 838 62100; fax: ⫹49 30 838 62620. E-mail address: [email protected](W. Heuwieser). 0093-691X/$ – see front matter © 2012 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.theriogenology.2012.07.024
Typically, fever is being defined as a BT exceeding a predefined value. Temperature thresholds of ⱖ 39.4 °C and ⱖ 39.7 °C [1–3] have been recommended to distinguish between healthy cows and cows suffering from an infectious disease. Monitoring rectal temperature (RT) for 5 to 10 days after parturition has received remarkable attention in the past two decades because of its cost-effectiveness as well as reliability [4] and has been incorporated as a routine diagnostic component
2032
O. Burfeind et al. / Theriogenology 78 (2012) 2031–2038
into protocols for early postpartum cow management [1]. In such protocols fever plays an important role in the daily diagnosis of infectious diseases, such as puerperal metritis [5]. However, there are several factors influencing the result of measurements of RT, such as the diurnal rhythm [6], the measuring process [7], or environmental conditions [8]. Recently an effort was made to determine the normal RT range in dairy cattle during the first 10 days in milk (DIM) for both healthy primiparous (N ⫽ 79) and healthy multiparous (N ⫽ 118) cows [9] measuring RT once daily in the morning. A temperature range from 37.9 °C to 39.6 °C (mean 38.8 °C) and 37.9 °C to 39.5 °C (mean 38.7 °C) was determined for primiparous and multiparous cows, respectively by calculating the mean RT minus and plus twice the standard deviation. In this study, however, classification of cows as healthy was based on a mere visual examination. Only cows that appeared abnormal were examined more closely. Therefore, sick cows with ambiguous signs of disease might have been misclassified as healthy and thus confounded the temperature range determined. In another study a mean RT ranging from 38.5 °C to 38.8 °C with a maximum from 38.9 °C to 39.4 °C was determined for the first 10 DIM in healthy cows (N ⫽ 569) with no differences between primiparous and multiparous cows [10]. Data for this study were retrospectively generated from on farm records documented by farm personnel which again might have led to a misclassification of undetected sick cows and biased results. A more intensive approach to diagnose diseases was performed in another study that determined temperature ranges for healthy dairy cows during the first 10 DIM [3]. In this study cows were examined for signs of diseases every 4 h during the first 10 DIM and cows that showed any clinical signs were classified as diseased. Although only 35 healthy cows were studied the reported mean RT was in accordance with the two studies stated above (38.7 °C to 38.8 °C). Approaches to determine physiological reference ranges for BT are challenged by the fact that BT is an integral part of the health status or the disease definition, respectively. In studies treating BT as a dependent variable the same parameter cannot be used to distinguish between healthy and diseased cows. Diagnosing a cow as suffering from puerperal metritis without reference to BT complicates the diagnosis enormously. To determine the false positive rate of a diagnostic method considering a certain threshold an independent reference has to be used [11]. In order to study false positive rates of fever in healthy cows BT has to be excluded
and another reliable parameter used for the definition of the health status ensuring that the cows selected for analysis were not suffering from infectious diseases. It is well known that haptoglobin, an acute phase protein, is increased in cows suffering from puerperal metritis [12,13]. Therefore, in this study we used serum haptoglobin concentration as inclusion criteria for the cows enrolled to ensure they were not suffering from puerperal metritis. In most studies investigating the early postpartum period RT was measured only once a day [2,4,14] with digital thermometers. Only two studies described frequent measurements of BT for early postpartum dairy cows [6,8]. Elevated BT values, however, are an important component for plausible treatment decisions regarding antibiotic (systemic or intrauterine) drugs alone or in combination with supportive therapy (antiinflammatory drugs, fluids). Frequency of type I errors (i.e., fever when the animal is actually healthy) was considerable (14% to 66%) even under moderate weather conditions [3,9,10]. It is well documented that changes in ambient temperature (AT) affect BT and activate thermoregulatory responses to maintain thermostability [15]. Furthermore, it has been demonstrated that BT is influenced earlier than other parameters, such as sweating or reduced dry matter intake [15–17]. The most common index used to define heat stress in dairy cattle is temperature humidity index (THI) computed from AT and relative humidity (RH). Several studies used THI as a measure of thermal comfort with a THI ⱖ 72 (equivalent to 25 °C AT and 50% RH) generally accepted as the upper threshold of the comfort zone for cattle [18 – 20]. Therefore, the overall objective of our study was to determine normal BT ranges in healthy dairy cows postpartum considering moderate and hot climate conditions. Specifically, we set out (1) to determine reference ranges for both RT and vaginal temperature (VT), (2) to evaluate the prevalence of fever considering three different thresholds (ⱖ 39.5 °C, ⱖ 39.7 °C, and ⱖ 40.0 °C) for RT and VT under moderate and hot weather conditions, and (3) to evaluate relationships between environmental parameters (AT, RH, THI) and BT under moderate and hot weather conditions in the early postpartum period. 2. Materials and methods From May to July 2010 an observational study on BT in the first 10 days of lactation of Holstein dairy cows was conducted on a commercial dairy farm in
O. Burfeind et al. / Theriogenology 78 (2012) 2031–2038
Sachsen-Anhalt, Germany. The herd consisted of 1200 dairy cows with an average 305 days milk production of 10 124 kg (4.1% fat and 3.4% protein). Cows were managed according to the guidelines set by the International Cooperation on Harmonisation of Technical Requirements for Registration of Veterinary Medicinal Products [21]. All cows were housed indoor in a freestall facility with slotted floors and cubicles equipped with rubber mats. Group composition was dynamic with cows entering and leaving the experiment depending on their calving dates. Early postpartum cows were fed a total mixed ration consisting of 42.7% concentrate and mineral mix, 32.2% corn silage, 22.6% grass silage, and 2.5% barley straw on a dry matter basis (net energy for lactation ⫽ 7.03 MJ/kg dry matter) distributed with a conveyor belt system up to 10 times a day. Cows were milked three times a day (approximately at 06.00 h, 15.00 h, and 22.00 h). Rectal temperature was measured twice daily from Day 2 to 10 after parturition by the study personnel (approximately 07.00 ⫾ 1 h and 17.00 ⫾ 1 h) using a digital thermometer (VT 1831, Microlife AG, Widnau, Switzerland) at an insertion depth of 8 cm. The digital thermometer measured RT from 34 °C to 42 °C with an accuracy of ⫾ 0.1 °C and a resolution of 0.1 °C. Vaginal temperature was monitored with a microprocessor controlled temperature logger (Minilog 8, Vemco, Ltd., Halifax, Nova Scotia, Canada) attached to a modified vaginal controlled internal drug release as recently validated [6]. Temperature loggers (size ⫽ 92 ⫻ 20 mm; weight ⫽ 40.5 g) were inserted in the vaginal cavity for an 8-day period from 2 to 10 days after parturition and measured VT every 10 min. The vaginal loggers measured VT from 0 to ⫹ 42 °C with an accuracy of ⫾ 0.3 °C and a resolution of 0.2 °C. Furthermore, on DIM 2, 5, and 10, blood samples were collected from all cows by puncture of coccygeal vessels using vacuum tubes (Vacuette, Greiner Bio-one, GmbH, Kremsmünster, Austria). Blood samples were centrifuged at 3500 X g for 10 min and two aliquots of the serum stored at ⫺25 °C until analysis. Serum haptoglobin concentrations were analyzed using an ELISA (Sunrice Reader, Tecan, Maennedorf, Switzerland) in a commercial laboratory (Synlab, Berlin, Germany). The intra- and interassay CV were 5.3% to 6.3% and 4.1% to 5.7%, respectively. Ambient temperature and RH within the experimental barn were recorded every 30 min using two Tinytag Plus II loggers (Gemini Loggers, Ltd., Chichester, UK) that were secured at beams 3 m from the ground 40 m apart. These loggers measured AT from ⫺25 °C to ⫹85 °C with an accuracy of ⫾ 0.3 °C and a resolution of 0.01
2033
°C and RH from 0 to 100% with an accuracy of ⫾ 3% and a resolution of 0.3%. Temperature humidity index was calculated according to the equation reported by Kendall et al. [22]: THI ⫽ (1.8 ⫻ AT ⫹ 32) ⫺ [(0.55 ⫺ 0.0055 ⫻ RH) ⫻ (1.8 ⫻ AT ⫺ 26)]. To study the effect of heat stress on BT, two weather periods with distinctly different THI were compared. During the first (May 1 to May 29, 2010) and second (June 23 to July 19, 2010) period the climate was moderate (THI 59.8 ⫾ 3.8) and hot (THI 74.1 ⫾ 4.4), respectively. Within these periods a total of 34 multiparous cows that had calved spontaneously (i.e., no assistance), delivered a single live calf, shedded the placenta within 12 h and did not receive any treatment (i.e., antibiotic or anti-inflammatory drugs, fluid therapy) during the observational period of 10 days were selected. To ensure that none of these cows was suffering from inflammation (i.e., puerperal metritis) the mean serum haptoglobin concentration was calculated from DIM 2, 5, and 10. Only cows with a mean serum haptoglobin concentration ⬍ 1.1 g/L that calved within the two different periods were selected for analysis to ensure their uterine health [13]. Because of serum habtoglobin concentrations exceeding this threshold two cows from the hot period were excluded from analysis. Therefore, a total of 17 and 15 cows met the inclusion criteria and were used for the final analyses during the moderate and hot period, respectively. Data from VT and THI loggers were downloaded into Excel spreadsheets (Office 2003, Microsoft Deutschland, GmbH, Munich, Germany) and analyzed using SPSS for Windows (Version 19.0, IBM Deutschland, GmbH, Ehningen, Germany). Vaginal temperature values below 38.0 °C were considered artifacts because of confounding by AT after logger movements. Hourly AT, RH, and THI means were produced averaging a total of four measures (i.e., two from each THI logger) for each variable during the whole experiment. Hourly VT means for every individual cow were produced averaging six VT measurements. Relationship and difference between RT and VT was determined using Pearson’s correlation and a paired t test. Relationship between RT and VT with environmental parameters (AT, RH, THI) was determined using Pearson’s correlation for both periods (moderate and hot). The proportion of THI exceeding 72 during both periods was compared using chi-square test. During the two weather periods the proportion of THI exceeding 72 was significantly different and therefore period was used as an independent factor with a nominal scale (moderate and hot). The influence of
2034
O. Burfeind et al. / Theriogenology 78 (2012) 2031–2038
Table 1 Association of rectal and vaginal temperature of healthy postpartum (days in milk 2 to 10) dairy cows with environmental parameters (ambient temperature, relative humidity, temperature humidity index) during moderate (temperature humidity index: 59.8 ⫾ 3.8) and hot climate (temperature humidity index: 74.1 ⫾ 4.4). Temperature Rectal
Vaginal
Parameters Ambient temperature (°C) Relative humidity (%) Temperature humidity index Ambient temperature (°C) Relative humidity (%) Temperature humidity index
Moderate period
Hot period
N*
r (P value)
N
271 271 271 3260 3260 3260
0.045 (0.463) ⫺0.04 (0.564) 0.05 (0.463) 0.10 (⬍0.001) ⫺0.10 (0.7) 0.01 (⬍0.001)
239 239 239 3231 3231 3231
r (P value) 0.46 (⬍0.001) ⫺0.27 (⬍0.001) 0.46 (⬍0.001) 0.34 (⬍0.001) ⫺0.10 (⬍0.001) 0.40 (⬍0.001)
* Number of paired observations.
biologically plausible factors (time of day, DIM, period, milk yield) and RT and VT, respectively, were tested repeating DIM in a mixed model procedure. The random effect of cows within the two weather periods was adjusted in both models (RT and VT). The scaled identity structure was used, because it resulted in the model with the lowest Aikaike information criterion value. First, the influence of each independent variable on RT and VT, respectively, was tested in a univariate analysis. Only variables with a P value ⬍ 0.20 were included in the final mixed models for RT and VT, respectively. In both models (RT and VT) milk yield was excluded because of lack of relationship to RT (P ⫽ 0.55) and VT (P ⫽ 0.22), respectively. The remaining variables (time of day, DIM, period) were included in the mixed model which was constructed in a manual backward stepwise manner. Therefore, variables with a P value ⬎ 0.05 were excluded from the model. Potential interactions of the final variables were tested in both models and kept if the P value was ⬍ 0.05. The prevalence of fever using three different thresholds (ⱖ 39.5 °C, ⱖ 39.7 °C, ⱖ 40.0 °C) was calculated for the daily RT and VT for both periods. Furthermore, it was calculated for a time from 07.00 h to 11.00 h in VT (in RT the morning measurement was used), a timeframe during which it is common to measure BT for producers. For both approaches the prevalence of fever was compared between both periods for RT and VT using chi-square test, respectively. To describe ranges of RT and VT of healthy cows during the first 10 DIM the mean temperature and the 95% confidence interval (CI95, RT, and VT) was calculated for both periods (moderate and hot). 3. Results A total of 19 449 and 17 204 ten-minute VT values were recorded for the moderate (N ⫽ 17 cows) and the
hot (N ⫽ 15 cows) period, respectively. After exclusion of 75 (0.04%) and 100 (0.06%) observations because of erroneous (⬍ 38 °C) measurements 3260 and 3231 hourly VT means were calculated for the moderate and hot period, respectively. Rectal and vaginal temperature were correlated (r ⫽ 0.85, N ⫽ 507 paired observations, P ⬍ 0.001), however, VT was 0.2 ⫾ 0.3 °C higher than RT (P ⬍ 0.001). The associations between RT and VT and environmental parameters (AT, RH, THI) were higher in the hot period than in the moderate period as presented in Table 1. Day in milk (2 to 10, P ⬍ 0.01), period (moderate and hot, P ⬍ 0.01), time of day (07.00 ⫾ 1 and 17.00 ⫾ 1, P ⬍ 0.01) and the interaction of period and time of day had an effect on RT (P ⬍ 0.01, Fig. 1). Vaginal temperature was influenced by period (moderate and hot, P ⬍ 0.01), time of day (24 h, P ⬍ 0.01), and DIM (2 to 10, P ⬍ 0.01, Figs. 2 and 3). The interactions of DIM and period influenced VT (Fig. 2, P ⬍ 0.01). Furthermore, the interaction of time of day and period influenced VT (Fig. 3, P ⬍ 0.01). The prevalence of fever defined by three different thresholds was higher in the hot than in the moderate period considering all values measured in the two periods (Table 2). Similarly, the prevalence of fever was higher in the hot than in the moderate period considering only the measures of RT in the morning or the measures of VT during a period from 07.00 h to 11.00 h, a time of day when fever is commonly measured by producers (Table 3). During the moderate period mean RT was 38.8 ⫾ 0.4 °C with the CI95 ranging from 38.0 °C to 39.7 °C in the morning (153 measurements from 17 cows) and 38.9 ⫾ 0.4 °C with the CI95 ranging from 38.1 °C to 39.7 °C in the evening (153 measurements from 17 cows). Overall mean RT was 38.9 ⫾ 0.4 °C with the CI95 ranging from 38.0 °C to 39.7 °C (306 measure-
O. Burfeind et al. / Theriogenology 78 (2012) 2031–2038
Fig. 1. Rectal temperature of healthy postpartum (days in milk 2 to 10) dairy cows during a moderate (N ⫽ 17, dotted line; temperature humidity index: 59.8 ⫾ 3.8) and a hot period (N ⫽ 15, solid line; temperature humidity index: 74.1 ⫾ 4.4).
ments of 17 cows). During the hot period mean RT was 39.2 ⫾ 0.5 °C with the CI95 ranging from 38.2 °C to 40.2 °C in the morning (135 measurements from 15 cows) and 39.6 ⫾ 0.6 °C with the CI95 ranging from 38.4 °C to 40.7 °C in the evening (135 measurements from 15 cows). Overall mean RT was 39.4 ⫾ 0.6 °C with the CI95 ranging from 38.3 °C to 40.3 °C (270 measurements from 15 cows). During the moderate period mean VT was 39.2 ⫾ 0.3 °C with the CI95 ranging from 38.5 °C to 39.8 °C (19 874 measurements from 17 cows). During the hot period mean VT was 39.6 ⫾ 0.4 °C with the CI95 ranging from 38.8 °C to 40.4 °C (17 104 measurements from 15 cows). Mean and CI95 of RT and VT considering DIM are presented in Table 4. For RT it is presented separately for the measurement in the morning and in the evening. For VT the daily mean is presented.
2035
plays an important role in the definition of health, especially for the absence of infectious diseases, such as puerperal metritis and mastitis [1]. Recent work that determined BT ranges of healthy cows classified cows as healthy based on a clinical examination alone in prospective studies [3,9] or retrospectively based on on-farm records [10]. In these studies BT itself was part of the definition of sickness. To determine false positive rate of a diagnostic method considering a certain threshold, however, an independent reference has to be used [11]. Therefore, in this study we used the serum haptoglobin concentration as an inclusion criterion for the cows studied to ensure they were not suffering from puerperal metritis [12,13]. Only cows with a mean serum haptoglobin concentration ⱕ 1.1 g/L were used for analysis. Using this criterion allowed excluding BT (RT and VT) from the definition of the health status (i.e., healthy or metritic), an important requirement when the false positive rate of fever in healthy cows is the variable of interest. The CI95 of RT determined in the moderate period (38.0 °C to 39.7 °C; Table 4) was in accordance with that reported by Wenz et al. [9] for 118 multiparous cows (37.9 °C to 39.5 °C). However, the CI95 determined in our study should be interpreted with caution, because they are only based on 17 (moderate) and 15 (hot) measurements for each DIM, re-
4. Discussion The main objective of this study was to establish reference ranges for BT (RT and VT) of healthy dairy cows under moderate and hot weather conditions and to determine the false positive rate of fever. In most studies evaluating efficacy of interventions and fresh cow protocols implemented on commercial dairy farms BT
Fig. 2. Vaginal temperature of healthy postpartum (days in milk 2 to 10) dairy cows during a moderate (N ⫽ 17, dotted line; temperature humidity index: 59.8 ⫾ 3.8) and a hot period (N ⫽ 15, solid line; temperature humidity index: 74.1 ⫾ 4.4).
2036
O. Burfeind et al. / Theriogenology 78 (2012) 2031–2038 Table 3 Difference of the prevalence of fever defined by three different thresholds (ⱖ 39.5 °C, ⱖ 39.7 °C, and ⱖ 40.0 °C) measuring rectal in the morning (07.00 ⫾ 1 h) and vaginal temperature from 07.00 h to 11.00 h between the moderate (temperature humidity index: 59.8 ⫾ 3.8) and hot period (temperature humidity index: 74.1 ⫾ 4.4). Temperature
Rectal (07.00 ⫾ 1 h)
Vaginal (07.00 to 11.00 h)
Fig. 3. Daily vaginal temperature rhythm (mean ⫾ SD) of healthy postpartum (days in milk 2 to 10) dairy cows during a moderate (N ⫽ 17, dotted line; temperature humidity index: 59.8 ⫾ 3.8) and a hot period (N ⫽ 15, solid line; temperature humidity index: 74.1 ⫾ 4.4).
spectively. Nevertheless, RT and VT are highly correlated with a constant difference of 0.2 °C and a similar CI95. This provides additional evidence for the RT ranges (CI95) calculated for the different DIM. Mean RT in the moderate period (38.9 ⫾ 0.4°C) was slightly higher than that reported by others (38.7 °C [9], 38.5 °C to 38.8 °C [10], 38.7 °C to 38.8 °C [3]). During the hot period mean RT was higher (39.4 ⫾ 0.6 °C) than the ones reported above for moderate conditions, which is very likely to be due to the greater impact of THI on BT
Table 2 Difference of the prevalence of fever defined by three different thresholds (ⱖ 39.5 °C, ⱖ 39.7 °C, and ⱖ 40.0 °C) measuring rectal (07.00 ⫾ 1 h and 17.00 ⫾ 1 h) and continuously (00.00 h to 24.00 h) measured vaginal temperature between the moderate (temperature humidity index: 59.8 ⫾ 3.8) and hot period (temperature humidity index: 74.1 ⫾ 4.4). Temperature
Rectal
Vaginal
Threshold (°C) ⱖ39.5 ⱖ39.7 ⱖ40 ⱖ39.5 ⱖ39.7 ⱖ40
Periods Moderate (%)
Hot (%)
P value
7.6 5.3 3.4 16.5 9.3 4.0
46.6 36.1 17.1 54.1 42.6 23.7
⬍0.001 ⬍0.001 ⬍0.001 ⬍0.001 ⬍0.001 ⬍0.001
Threshold (°C) ⱖ39.5 ⱖ39.7 ⱖ40 ⱖ39.5 ⱖ39.7 ⱖ40
Periods Moderate (%)
Hot (%)
P value
7.4 4.1 2.7 10.0 5.9 3.3
28.1 21.4 8.8 33.0 21.2 8.6
⬍0.001 ⬍0.001 0.02 ⬍0.001 ⬍0.001 ⬍0.001
under hot environmental conditions than under moderate ones [20]. This hypothesis is supported by the higher correlation of THI and RT and VT, respectively, during the hot period (RT: r ⫽ 0.46; VT: r ⫽ 0.40) compared with the moderate period (RT: r ⫽ 0.05; VT: r ⫽ 0.01). Furthermore, in the three studies stated above no systematic continuous measurement of the climatic condition (AT, RH, THI) prevailing on the farm was performed. To our knowledge this is the first study to monitor THI and quantify effects of hot weather conditions on physiological BT ranges. Mean VT during the moderate period in this study was higher (39.2 ⫾ 0.3 °C) than mean RT in this study (38.9 ⫾ 0.4 °C) and higher than the results from the studies cited above. Furthermore, mean VT during the hot period (39.6 ⫾ 0.4 °C) was higher than RT during the hot period (39.4 ⫾ 0.6 °C). This phenomena has been described before [23] and might be due to the fact that the logger is deeper in the body when inserted into the vagina than the probe of a thermometer inserted in the rectum. In this study we determined the prevalence of fever in two different approaches. In the first approach we averaged the two measures of RT and the 24 hourly measures of VT, respectively, to one daily RT and VT and described the percentage of measures exceeding three different thresholds. This approach is comparable to Wagner et al. [3] who measured RT every 4 h and determined the prevalence of fever (i.e., at least one measurement reaching the defined threshold during the first 10 DIM) of 66% (39.5 °C), 49% (39.7 °C), and 17% (40.0 °C). Interestingly these prevalence rates are much higher than the ones estimated in this study during the moderate period, but comparable with the ones estimated during the hot period (Table 2). However, the
O. Burfeind et al. / Theriogenology 78 (2012) 2031–2038
2037
Table 4 Mean rectal and vaginal temperature of healthy dairy cows during the first 10 days in milk (DIM) during a moderate (N ⫽ 17 cows; temperature humidity index: 59.8 ⫾ 3.8) and a hot period (N ⫽ 15 cows; temperature humidity index: 74.1 ⫾ 4.4). Period
Moderate
Hot
DIM
2 3 4 5 6 7 8 9 10 2 3 4 5 6 7 8 9 10
Rectal temperature (°C), mean (95% confidence interval) Morning (07.00 h)
Evening (17.00 h)
38.6 (37.9–39.2) 38.6 (38.1–39.2) 38.7 (38.1–39.4) 38.8 (38.1–39.6) 38.9 (38.2–39.5) 38.9 (37.9–40.0) 39.0 (38.0–40.0) 39.0 (37.7–40.2) 38.8 (38.2–39.5) 38.9 (38.0–39.6) 39.1 (38.3–39.8) 39.1 (38.1–40.0) 39.3 (38.2–40.4) 39.4 (38.4–40.4) 39.4 (38.3–40.6) 39.3 (38.5–40.1) 39.4 (38.6–40.2) 39.1 (38.2–40.0)
38.8 (38.1–39.5) 38.7 (38.0–39.4) 39.0 (37.9–40.1) 39.0 (38.2–39.8) 38.9 (38.1–39.7) 38.8 (38.1–39.6) 38.9 (37.8–40.0) 38.8 (38.3–39.3) 38.9 (38.3–39.6) 39.6 (38.3–40.8) 39.6 (38.1–41.1) 39.5 (38.3–40.7) 39.6 (38.5–40.8) 39.6 (38.4–40.7) 39.7 (38.5–40.9) 39.5 (38.6–40.4) 39.5 (38.3–40.7) 39.5 (38.5–40.5)
study of Wagner et al. [3] was conducted from fall to spring (September 2005 to April 2006) in North Dakota (USA), a region with continental climate. Therefore, the prevalence should be compared with the ones during the moderate period and not the hot period in our study. A possible explanation could be that health was determined by a clinical examination [3] and not by serum haptoglobin concentrations as in our study. Therefore, a possible misclassification of sick cows as healthy might have increased the prevalence of fever. In the second approach we determined the prevalence of fever for RT based on the measurement conducted in the morning (07.00 ⫾ 1 h) and for VT based on the average temperature from 07.00 h to 11.00 h. This approach was simulating a common routine of herdsmen who measure RT in the morning hours. Also this protocol was comparable to Wenz et al. [9] who reported a prevalence of fever (i.e., at least one measurement above 39.5 °C during the first 10 DIM) of 15% for multiparous cows. This is slightly higher than the prevalence in this study calculated for the moderate period, but lower than for the hot period (Table 3). The study was conducted in Colorado (USA), from May to July 2005 and 2006. The AT ranged from 7.2 °C to 25 °C, and most likely included moderate and hot periods. The THI, however, was not calculated. This might be a possible explanation for the estimated prevalence of fever lying in between the ones estimated in our study during the moderate and the hot period. Furthermore, in
Vaginal temperature (°C), mean (95% confidence interval) 38.9 (38.3–39.5) 38.9 (38.3–39.6) 39.1 (38.3–39.9) 39.2 (38.3–40.1) 39.2 (38.5–40.0) 39.2 (38.5–40.0) 39.3 (38.4–40.2) 39.2 (38.4–40.0) 39.2 (38.6–39.8) 39.5 (38.3–40.7) 39.5 (38.4–40.6) 39.5 (38.5–40.5) 39.6 (38.5–40.7) 39.7 (38.6–40.9) 39.7 (38.6–40.9) 39.6 (38.7–40.6) 39.6 (38.6–40.6) 39.5 (38.5–40.5)
the study conducted by Wenz et al. [9] cows were defined as healthy based on the production group (i.e., infectious disease, metabolic/digestive disease, no disease) to which they belonged. While health events were identified by farm personnel, metritis was diagnosed by farm and study personnel but with different definitions of metritis with respect to RT. Cows identified with metritis by study personnel did not necessarily have an RT ⬎ 39.4 °C, whereas this was a necessary requirement in the diagnosis of metritis by farm personnel. Therefore, 34% (50 of 145 cows) that were diagnosed with metritis by study personnel but not by farm personnel did not received any antibiotic treatment and were considered healthy. The authors speculated that these cows might have had less severe disease and therefore were not detected by farm personnel [9]. However, this fact clearly illustrates the need of a more specific parameter to define cows as healthy (i.e., serum haptoglobin concentration) in a research setting studying BT in healthy dairy cows. In this study RT and VT were influenced by several factors (DIM, period, time of day). For RT an additional interaction of DIM and time of day was present, whereas two additional interactions (DIM and period, time of day and period) influenced VT. Wenz et al. [9] identified similar factors influencing RT (month of calving, DIM, parity, disease group), including primiparous cows and diseased cows in their analyses. Our intention, however, was to describe RT and VT in
2038
O. Burfeind et al. / Theriogenology 78 (2012) 2031–2038
healthy cows, defined by a serum haptoglobin concentration ⱕ 1.1 g/L.
[7]
4.1. Conclusions This study demonstrates that BT 10 days postpartum is influenced by several factors (DIM, climate, time of day). Therefore, these factors have to be considered when interpreting BT measures to identify sick cows. Furthermore, the prevalence of fever considering different thresholds is higher during hot than moderate periods. However, even in a moderate period healthy cows can exhibit BT that is considered as fever. The data support that to minimize false positives, RT should be routinely taken in the morning. During moderate periods (THI 59.8 ⫾ 3.8), 7.4% of cows exhibited RT ⱖ 39.5 °C and during hot period (THI 74.1 ⫾ 4.4), 28.1% of cows exhibited RT ⱖ 39.5°C. This fact clearly illustrates that fever alone should not be considered the decision criterion whether a cow is allocated to an antibiotic treatment, although it is the most important one that is objectively measurable.
[8]
[9]
[10]
[11] [12]
[13]
[14]
Acknowledgments
[15]
The authors thank the owners, the herdswoman, and the staff of the farm. This project was funded in part by Pfizer Animal Health, Veterinary Medicine Research and Development, UK.
[16] [17]
[18]
References [1] Smith BI, Risco CA. Management of periparturient disorders in dairy cattle. Vet Clin North Am Food Anim Pract 2005;21:503– 21. [2] Benzaquen ME, Risco CA, Archbald LF, Melendez P, Thatcher MJ, Thatcher WW. Rectal temperature, calving-related factors, and the incidence of puerperal metritis in postpartum dairy cows. J Dairy Sci 2007;90:2804 –14. [3] Wagner SA, Schimeck DE, Chend FC. Body temperature and white blood cell count in postpartum dairy cows. The Bovine Practitioner 2008;42:18 –26. [4] Zhou C, Boucher JF, Dame KJ, Moreira M, Graham R, Nantel J, et al. Multilocation trial of ceftiofur for treatment of postpartum cows with fever. J Am Vet Med Assoc 2001;219:805– 8. [5] Sheldon IM, Lewis GS, LeBlanc S, Gilbert RO. Defining postpartum uterine disease in cattle. Theriogenology 2006;65:1516 –30. [6] Vickers LA, Burfeind O, von Keyserlingk MA, Veira DM, Weary DM, Heuwieser W. Technical note: comparison of rectal
[19]
[20]
[21]
[22]
[23]
and vaginal temperatures in lactating dairy cows. J Dairy Sci 2010;93:5246 –51. Burfeind O, von Keyserlingk MA, Weary DM, Veira DM, Heuwieser W. Short communication: repeatability of measures of rectal temperature in dairy cows. J Dairy Sci 2010;93:624 –7. Suthar V, Burfeind O, Bonk S, Voigtsberger R, Keane C, Heuwieser W. Factors associated with body temperature of healthy holstein dairy cows during the first 10 days in milk. J Dairy Res 2012;79:135– 42. Wenz JR, Moore DA, Kasimanickam R. Factors associated with the rectal temperature of holstein dairy cows during the first 10 days in milk. J Dairy Sci 2011;94:1864 –72. Kristula M, Smith AM, Simeone A. The use of daily postpartum rectal temperatures to select dairy cows for treatment with systemic antibiotics. The Bovine Practitioner 2001;35:117–25. Martin SW. The evaluation of tests. Can J Comp Med 1977;41: 19 –25. Huzzey JM, Duffield TF, LeBlanc SJ, Veira DM, Weary DM, von Keyserlingk MA. Short communication: haptoglobin as an early indicator of metritis. J Dairy Sci 2009;92:621–5. Drillich M, Voigt D, Forderung D, Heuwieser W. Treatment of acute puerperal metritis with flunixin meglumine in addition to antibiotic treatment. J Dairy Sci 2007;90:3758 – 63. Drillich M, Beetz O, Pfützner A, Sabin M, Sabin HJ, Kutzer P, et al. Evaluation of a systemic antibiotic treatment of toxic puerperal metritis in dairy cows. J Dairy Sci 2001;84:2010 –7. Kadzere CT, Murphy MR, Silanikove N, Maltz E. Heat stress in lactating dairy cows: a review. Livest Prod Sci 2002;77:59 –91. Hahn GL. Dynamic responses of cattle to thermal heat loads. J Anim Sci 1999;77 Suppl 2:10 –20. Silanikove N. Effects of heat stress on the welfare of extensively managed domestic ruminants. Livest Prod Sci 2000;67: 1–18. Igono MO, Bjotvedt G, Sanford-Crane HT. Environmental profile and critical temperature effects on milk production of holstein cows in desert climate. Int J Biometeorol 1992;36:77– 87. West JW, Mullinix BG, Bernard JK. Effects of hot, humid weather on milk temperature, dry matter intake, and milk yield of lactating dairy cows. J Dairy Sci 2003;86:232– 42. Kendall PE, Nielsen PP, Webster JR, Verkerk GA, Littlejohn RP, Matthews LR. The effects of providing shade to lactating dairy cows in a temperate climate. Livest Sci 2006;103:148 –57. Hellmann K, Radeloff I. International cooperation on harmonisation of technical requirements for registration of veterinary medicinal products. VICH, 2000. http://www.fda.gov/downloads/Animal Veterinary/GuidanceComplianceEnforcement/GuidanceforIndustry/ UCM052417.pdf. Accessed August 2012. Kendall PE, Tucker CB, Dalley DE, Clark DA, Webster JR. Milking frequency affects the circadian body temperature rhythm in dairy cows. Livest Sci 2008;117:130 – 8. Burfeind O, Suthar VS, Voigtsberger R, Bonk S, Heuwieser W. Validity of prepartum changes in vaginal and rectal temperature to predict calving in dairy cows. J Dairy Sci 2011;94:5053– 61.
Theriogenology 78 (2012) 2031–2038 www.theriojournal.com
Effect of heat stress on body temperature in healthy early postpartum dairy cows O. Burfeind, V.S. Suthar, W. Heuwieser* Clinic of Animal Reproduction, Faculty of Veterinary Medicine, Freie Universität Berlin, Berlin, Germany
Abstract Measurement of body temperature is the most common method for an early diagnosis of sick cows in fresh cow protocols currently used on dairy farms. Thresholds for fever range from 39.4 °C to 39.7 °C. Several studies attempted to describe normal temperature ranges for healthy dairy cows in the early puerperium. However, the definition of a healthy cow is variable within these studies. It is challenging to determine normal temperature ranges for healthy cows because body temperature is usually included in the definition. Therefore, the objectives of this study were to identify factors that influence body temperature in healthy dairy cows early postpartum and to determine normal temperature ranges for healthy cows that calved in a moderate (temperature humidity index: 59.8 ⫾ 3.8) and a hot period (temperature humidity index: 74.1 ⫾ 4.4), respectively, excluding body temperature from the definition of the health status. Furthermore, the prevalence of fever was calculated for both periods separately. A subset of 17 (moderate period) and 15 cows (hot period) were used for analysis. To ensure their uterine health only cows with a serum haptoglobin concentration ⱕ 1.1 g/L were included in the analysis. Therefore, body temperature could be excluded from the definition. A vaginal temperature logger that measured vaginal temperature every 10 min was inserted from Day 2 to 10 after parturition. Additionally rectal temperature was measured twice daily. Day in milk (2 to 10), period (moderate and hot), and time of day had an effect on rectal and vaginal temperature. The prevalence of fever (ⱖ 39.5 °C) was 7.4% and 28.1% for rectal temperature in the moderate and hot period, respectively. For vaginal temperature (07.00 to 11.00 h) it was 10% and 33%, respectively, considering the same threshold and period. This study demonstrates that body temperature in the early puerperium is influenced by several factors (day in milk, climate, time of day). Therefore, these factors have to be considered when interpreting body temperature measures to identify sick cows. Furthermore, the prevalence of fever considering different thresholds is higher during hot than moderate periods. However, even in a moderate period healthy cows can exhibit a body temperature that is considered as fever. This fact clearly illustrates that fever alone should not be considered the decision criterion whether a cow is allocated to an antibiotic treatment, although it is the most important one that is objectively measurable. © 2012 Elsevier Inc. All rights reserved. Keywords: Dairy cow; Body temperature; Fever; Climate; Haptoglobin; Fresh cow protocol
1. Introduction Currently, measurement of body temperature (BT) is one of the most common methods used for the diagnosis of infectious diseases (i.e., puerperal metritis, clinical mastitis) in the early puerperium in dairy cows [1]. * Corresponding author. Tel.: ⫹49 30 838 62100; fax: ⫹49 30 838 62620. E-mail address: [email protected](W. Heuwieser). 0093-691X/$ – see front matter © 2012 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.theriogenology.2012.07.024
Typically, fever is being defined as a BT exceeding a predefined value. Temperature thresholds of ⱖ 39.4 °C and ⱖ 39.7 °C [1–3] have been recommended to distinguish between healthy cows and cows suffering from an infectious disease. Monitoring rectal temperature (RT) for 5 to 10 days after parturition has received remarkable attention in the past two decades because of its cost-effectiveness as well as reliability [4] and has been incorporated as a routine diagnostic component
2032
O. Burfeind et al. / Theriogenology 78 (2012) 2031–2038
into protocols for early postpartum cow management [1]. In such protocols fever plays an important role in the daily diagnosis of infectious diseases, such as puerperal metritis [5]. However, there are several factors influencing the result of measurements of RT, such as the diurnal rhythm [6], the measuring process [7], or environmental conditions [8]. Recently an effort was made to determine the normal RT range in dairy cattle during the first 10 days in milk (DIM) for both healthy primiparous (N ⫽ 79) and healthy multiparous (N ⫽ 118) cows [9] measuring RT once daily in the morning. A temperature range from 37.9 °C to 39.6 °C (mean 38.8 °C) and 37.9 °C to 39.5 °C (mean 38.7 °C) was determined for primiparous and multiparous cows, respectively by calculating the mean RT minus and plus twice the standard deviation. In this study, however, classification of cows as healthy was based on a mere visual examination. Only cows that appeared abnormal were examined more closely. Therefore, sick cows with ambiguous signs of disease might have been misclassified as healthy and thus confounded the temperature range determined. In another study a mean RT ranging from 38.5 °C to 38.8 °C with a maximum from 38.9 °C to 39.4 °C was determined for the first 10 DIM in healthy cows (N ⫽ 569) with no differences between primiparous and multiparous cows [10]. Data for this study were retrospectively generated from on farm records documented by farm personnel which again might have led to a misclassification of undetected sick cows and biased results. A more intensive approach to diagnose diseases was performed in another study that determined temperature ranges for healthy dairy cows during the first 10 DIM [3]. In this study cows were examined for signs of diseases every 4 h during the first 10 DIM and cows that showed any clinical signs were classified as diseased. Although only 35 healthy cows were studied the reported mean RT was in accordance with the two studies stated above (38.7 °C to 38.8 °C). Approaches to determine physiological reference ranges for BT are challenged by the fact that BT is an integral part of the health status or the disease definition, respectively. In studies treating BT as a dependent variable the same parameter cannot be used to distinguish between healthy and diseased cows. Diagnosing a cow as suffering from puerperal metritis without reference to BT complicates the diagnosis enormously. To determine the false positive rate of a diagnostic method considering a certain threshold an independent reference has to be used [11]. In order to study false positive rates of fever in healthy cows BT has to be excluded
and another reliable parameter used for the definition of the health status ensuring that the cows selected for analysis were not suffering from infectious diseases. It is well known that haptoglobin, an acute phase protein, is increased in cows suffering from puerperal metritis [12,13]. Therefore, in this study we used serum haptoglobin concentration as inclusion criteria for the cows enrolled to ensure they were not suffering from puerperal metritis. In most studies investigating the early postpartum period RT was measured only once a day [2,4,14] with digital thermometers. Only two studies described frequent measurements of BT for early postpartum dairy cows [6,8]. Elevated BT values, however, are an important component for plausible treatment decisions regarding antibiotic (systemic or intrauterine) drugs alone or in combination with supportive therapy (antiinflammatory drugs, fluids). Frequency of type I errors (i.e., fever when the animal is actually healthy) was considerable (14% to 66%) even under moderate weather conditions [3,9,10]. It is well documented that changes in ambient temperature (AT) affect BT and activate thermoregulatory responses to maintain thermostability [15]. Furthermore, it has been demonstrated that BT is influenced earlier than other parameters, such as sweating or reduced dry matter intake [15–17]. The most common index used to define heat stress in dairy cattle is temperature humidity index (THI) computed from AT and relative humidity (RH). Several studies used THI as a measure of thermal comfort with a THI ⱖ 72 (equivalent to 25 °C AT and 50% RH) generally accepted as the upper threshold of the comfort zone for cattle [18 – 20]. Therefore, the overall objective of our study was to determine normal BT ranges in healthy dairy cows postpartum considering moderate and hot climate conditions. Specifically, we set out (1) to determine reference ranges for both RT and vaginal temperature (VT), (2) to evaluate the prevalence of fever considering three different thresholds (ⱖ 39.5 °C, ⱖ 39.7 °C, and ⱖ 40.0 °C) for RT and VT under moderate and hot weather conditions, and (3) to evaluate relationships between environmental parameters (AT, RH, THI) and BT under moderate and hot weather conditions in the early postpartum period. 2. Materials and methods From May to July 2010 an observational study on BT in the first 10 days of lactation of Holstein dairy cows was conducted on a commercial dairy farm in
O. Burfeind et al. / Theriogenology 78 (2012) 2031–2038
Sachsen-Anhalt, Germany. The herd consisted of 1200 dairy cows with an average 305 days milk production of 10 124 kg (4.1% fat and 3.4% protein). Cows were managed according to the guidelines set by the International Cooperation on Harmonisation of Technical Requirements for Registration of Veterinary Medicinal Products [21]. All cows were housed indoor in a freestall facility with slotted floors and cubicles equipped with rubber mats. Group composition was dynamic with cows entering and leaving the experiment depending on their calving dates. Early postpartum cows were fed a total mixed ration consisting of 42.7% concentrate and mineral mix, 32.2% corn silage, 22.6% grass silage, and 2.5% barley straw on a dry matter basis (net energy for lactation ⫽ 7.03 MJ/kg dry matter) distributed with a conveyor belt system up to 10 times a day. Cows were milked three times a day (approximately at 06.00 h, 15.00 h, and 22.00 h). Rectal temperature was measured twice daily from Day 2 to 10 after parturition by the study personnel (approximately 07.00 ⫾ 1 h and 17.00 ⫾ 1 h) using a digital thermometer (VT 1831, Microlife AG, Widnau, Switzerland) at an insertion depth of 8 cm. The digital thermometer measured RT from 34 °C to 42 °C with an accuracy of ⫾ 0.1 °C and a resolution of 0.1 °C. Vaginal temperature was monitored with a microprocessor controlled temperature logger (Minilog 8, Vemco, Ltd., Halifax, Nova Scotia, Canada) attached to a modified vaginal controlled internal drug release as recently validated [6]. Temperature loggers (size ⫽ 92 ⫻ 20 mm; weight ⫽ 40.5 g) were inserted in the vaginal cavity for an 8-day period from 2 to 10 days after parturition and measured VT every 10 min. The vaginal loggers measured VT from 0 to ⫹ 42 °C with an accuracy of ⫾ 0.3 °C and a resolution of 0.2 °C. Furthermore, on DIM 2, 5, and 10, blood samples were collected from all cows by puncture of coccygeal vessels using vacuum tubes (Vacuette, Greiner Bio-one, GmbH, Kremsmünster, Austria). Blood samples were centrifuged at 3500 X g for 10 min and two aliquots of the serum stored at ⫺25 °C until analysis. Serum haptoglobin concentrations were analyzed using an ELISA (Sunrice Reader, Tecan, Maennedorf, Switzerland) in a commercial laboratory (Synlab, Berlin, Germany). The intra- and interassay CV were 5.3% to 6.3% and 4.1% to 5.7%, respectively. Ambient temperature and RH within the experimental barn were recorded every 30 min using two Tinytag Plus II loggers (Gemini Loggers, Ltd., Chichester, UK) that were secured at beams 3 m from the ground 40 m apart. These loggers measured AT from ⫺25 °C to ⫹85 °C with an accuracy of ⫾ 0.3 °C and a resolution of 0.01
2033
°C and RH from 0 to 100% with an accuracy of ⫾ 3% and a resolution of 0.3%. Temperature humidity index was calculated according to the equation reported by Kendall et al. [22]: THI ⫽ (1.8 ⫻ AT ⫹ 32) ⫺ [(0.55 ⫺ 0.0055 ⫻ RH) ⫻ (1.8 ⫻ AT ⫺ 26)]. To study the effect of heat stress on BT, two weather periods with distinctly different THI were compared. During the first (May 1 to May 29, 2010) and second (June 23 to July 19, 2010) period the climate was moderate (THI 59.8 ⫾ 3.8) and hot (THI 74.1 ⫾ 4.4), respectively. Within these periods a total of 34 multiparous cows that had calved spontaneously (i.e., no assistance), delivered a single live calf, shedded the placenta within 12 h and did not receive any treatment (i.e., antibiotic or anti-inflammatory drugs, fluid therapy) during the observational period of 10 days were selected. To ensure that none of these cows was suffering from inflammation (i.e., puerperal metritis) the mean serum haptoglobin concentration was calculated from DIM 2, 5, and 10. Only cows with a mean serum haptoglobin concentration ⬍ 1.1 g/L that calved within the two different periods were selected for analysis to ensure their uterine health [13]. Because of serum habtoglobin concentrations exceeding this threshold two cows from the hot period were excluded from analysis. Therefore, a total of 17 and 15 cows met the inclusion criteria and were used for the final analyses during the moderate and hot period, respectively. Data from VT and THI loggers were downloaded into Excel spreadsheets (Office 2003, Microsoft Deutschland, GmbH, Munich, Germany) and analyzed using SPSS for Windows (Version 19.0, IBM Deutschland, GmbH, Ehningen, Germany). Vaginal temperature values below 38.0 °C were considered artifacts because of confounding by AT after logger movements. Hourly AT, RH, and THI means were produced averaging a total of four measures (i.e., two from each THI logger) for each variable during the whole experiment. Hourly VT means for every individual cow were produced averaging six VT measurements. Relationship and difference between RT and VT was determined using Pearson’s correlation and a paired t test. Relationship between RT and VT with environmental parameters (AT, RH, THI) was determined using Pearson’s correlation for both periods (moderate and hot). The proportion of THI exceeding 72 during both periods was compared using chi-square test. During the two weather periods the proportion of THI exceeding 72 was significantly different and therefore period was used as an independent factor with a nominal scale (moderate and hot). The influence of
2034
O. Burfeind et al. / Theriogenology 78 (2012) 2031–2038
Table 1 Association of rectal and vaginal temperature of healthy postpartum (days in milk 2 to 10) dairy cows with environmental parameters (ambient temperature, relative humidity, temperature humidity index) during moderate (temperature humidity index: 59.8 ⫾ 3.8) and hot climate (temperature humidity index: 74.1 ⫾ 4.4). Temperature Rectal
Vaginal
Parameters Ambient temperature (°C) Relative humidity (%) Temperature humidity index Ambient temperature (°C) Relative humidity (%) Temperature humidity index
Moderate period
Hot period
N*
r (P value)
N
271 271 271 3260 3260 3260
0.045 (0.463) ⫺0.04 (0.564) 0.05 (0.463) 0.10 (⬍0.001) ⫺0.10 (0.7) 0.01 (⬍0.001)
239 239 239 3231 3231 3231
r (P value) 0.46 (⬍0.001) ⫺0.27 (⬍0.001) 0.46 (⬍0.001) 0.34 (⬍0.001) ⫺0.10 (⬍0.001) 0.40 (⬍0.001)
* Number of paired observations.
biologically plausible factors (time of day, DIM, period, milk yield) and RT and VT, respectively, were tested repeating DIM in a mixed model procedure. The random effect of cows within the two weather periods was adjusted in both models (RT and VT). The scaled identity structure was used, because it resulted in the model with the lowest Aikaike information criterion value. First, the influence of each independent variable on RT and VT, respectively, was tested in a univariate analysis. Only variables with a P value ⬍ 0.20 were included in the final mixed models for RT and VT, respectively. In both models (RT and VT) milk yield was excluded because of lack of relationship to RT (P ⫽ 0.55) and VT (P ⫽ 0.22), respectively. The remaining variables (time of day, DIM, period) were included in the mixed model which was constructed in a manual backward stepwise manner. Therefore, variables with a P value ⬎ 0.05 were excluded from the model. Potential interactions of the final variables were tested in both models and kept if the P value was ⬍ 0.05. The prevalence of fever using three different thresholds (ⱖ 39.5 °C, ⱖ 39.7 °C, ⱖ 40.0 °C) was calculated for the daily RT and VT for both periods. Furthermore, it was calculated for a time from 07.00 h to 11.00 h in VT (in RT the morning measurement was used), a timeframe during which it is common to measure BT for producers. For both approaches the prevalence of fever was compared between both periods for RT and VT using chi-square test, respectively. To describe ranges of RT and VT of healthy cows during the first 10 DIM the mean temperature and the 95% confidence interval (CI95, RT, and VT) was calculated for both periods (moderate and hot). 3. Results A total of 19 449 and 17 204 ten-minute VT values were recorded for the moderate (N ⫽ 17 cows) and the
hot (N ⫽ 15 cows) period, respectively. After exclusion of 75 (0.04%) and 100 (0.06%) observations because of erroneous (⬍ 38 °C) measurements 3260 and 3231 hourly VT means were calculated for the moderate and hot period, respectively. Rectal and vaginal temperature were correlated (r ⫽ 0.85, N ⫽ 507 paired observations, P ⬍ 0.001), however, VT was 0.2 ⫾ 0.3 °C higher than RT (P ⬍ 0.001). The associations between RT and VT and environmental parameters (AT, RH, THI) were higher in the hot period than in the moderate period as presented in Table 1. Day in milk (2 to 10, P ⬍ 0.01), period (moderate and hot, P ⬍ 0.01), time of day (07.00 ⫾ 1 and 17.00 ⫾ 1, P ⬍ 0.01) and the interaction of period and time of day had an effect on RT (P ⬍ 0.01, Fig. 1). Vaginal temperature was influenced by period (moderate and hot, P ⬍ 0.01), time of day (24 h, P ⬍ 0.01), and DIM (2 to 10, P ⬍ 0.01, Figs. 2 and 3). The interactions of DIM and period influenced VT (Fig. 2, P ⬍ 0.01). Furthermore, the interaction of time of day and period influenced VT (Fig. 3, P ⬍ 0.01). The prevalence of fever defined by three different thresholds was higher in the hot than in the moderate period considering all values measured in the two periods (Table 2). Similarly, the prevalence of fever was higher in the hot than in the moderate period considering only the measures of RT in the morning or the measures of VT during a period from 07.00 h to 11.00 h, a time of day when fever is commonly measured by producers (Table 3). During the moderate period mean RT was 38.8 ⫾ 0.4 °C with the CI95 ranging from 38.0 °C to 39.7 °C in the morning (153 measurements from 17 cows) and 38.9 ⫾ 0.4 °C with the CI95 ranging from 38.1 °C to 39.7 °C in the evening (153 measurements from 17 cows). Overall mean RT was 38.9 ⫾ 0.4 °C with the CI95 ranging from 38.0 °C to 39.7 °C (306 measure-
O. Burfeind et al. / Theriogenology 78 (2012) 2031–2038
Fig. 1. Rectal temperature of healthy postpartum (days in milk 2 to 10) dairy cows during a moderate (N ⫽ 17, dotted line; temperature humidity index: 59.8 ⫾ 3.8) and a hot period (N ⫽ 15, solid line; temperature humidity index: 74.1 ⫾ 4.4).
ments of 17 cows). During the hot period mean RT was 39.2 ⫾ 0.5 °C with the CI95 ranging from 38.2 °C to 40.2 °C in the morning (135 measurements from 15 cows) and 39.6 ⫾ 0.6 °C with the CI95 ranging from 38.4 °C to 40.7 °C in the evening (135 measurements from 15 cows). Overall mean RT was 39.4 ⫾ 0.6 °C with the CI95 ranging from 38.3 °C to 40.3 °C (270 measurements from 15 cows). During the moderate period mean VT was 39.2 ⫾ 0.3 °C with the CI95 ranging from 38.5 °C to 39.8 °C (19 874 measurements from 17 cows). During the hot period mean VT was 39.6 ⫾ 0.4 °C with the CI95 ranging from 38.8 °C to 40.4 °C (17 104 measurements from 15 cows). Mean and CI95 of RT and VT considering DIM are presented in Table 4. For RT it is presented separately for the measurement in the morning and in the evening. For VT the daily mean is presented.
2035
plays an important role in the definition of health, especially for the absence of infectious diseases, such as puerperal metritis and mastitis [1]. Recent work that determined BT ranges of healthy cows classified cows as healthy based on a clinical examination alone in prospective studies [3,9] or retrospectively based on on-farm records [10]. In these studies BT itself was part of the definition of sickness. To determine false positive rate of a diagnostic method considering a certain threshold, however, an independent reference has to be used [11]. Therefore, in this study we used the serum haptoglobin concentration as an inclusion criterion for the cows studied to ensure they were not suffering from puerperal metritis [12,13]. Only cows with a mean serum haptoglobin concentration ⱕ 1.1 g/L were used for analysis. Using this criterion allowed excluding BT (RT and VT) from the definition of the health status (i.e., healthy or metritic), an important requirement when the false positive rate of fever in healthy cows is the variable of interest. The CI95 of RT determined in the moderate period (38.0 °C to 39.7 °C; Table 4) was in accordance with that reported by Wenz et al. [9] for 118 multiparous cows (37.9 °C to 39.5 °C). However, the CI95 determined in our study should be interpreted with caution, because they are only based on 17 (moderate) and 15 (hot) measurements for each DIM, re-
4. Discussion The main objective of this study was to establish reference ranges for BT (RT and VT) of healthy dairy cows under moderate and hot weather conditions and to determine the false positive rate of fever. In most studies evaluating efficacy of interventions and fresh cow protocols implemented on commercial dairy farms BT
Fig. 2. Vaginal temperature of healthy postpartum (days in milk 2 to 10) dairy cows during a moderate (N ⫽ 17, dotted line; temperature humidity index: 59.8 ⫾ 3.8) and a hot period (N ⫽ 15, solid line; temperature humidity index: 74.1 ⫾ 4.4).
2036
O. Burfeind et al. / Theriogenology 78 (2012) 2031–2038 Table 3 Difference of the prevalence of fever defined by three different thresholds (ⱖ 39.5 °C, ⱖ 39.7 °C, and ⱖ 40.0 °C) measuring rectal in the morning (07.00 ⫾ 1 h) and vaginal temperature from 07.00 h to 11.00 h between the moderate (temperature humidity index: 59.8 ⫾ 3.8) and hot period (temperature humidity index: 74.1 ⫾ 4.4). Temperature
Rectal (07.00 ⫾ 1 h)
Vaginal (07.00 to 11.00 h)
Fig. 3. Daily vaginal temperature rhythm (mean ⫾ SD) of healthy postpartum (days in milk 2 to 10) dairy cows during a moderate (N ⫽ 17, dotted line; temperature humidity index: 59.8 ⫾ 3.8) and a hot period (N ⫽ 15, solid line; temperature humidity index: 74.1 ⫾ 4.4).
spectively. Nevertheless, RT and VT are highly correlated with a constant difference of 0.2 °C and a similar CI95. This provides additional evidence for the RT ranges (CI95) calculated for the different DIM. Mean RT in the moderate period (38.9 ⫾ 0.4°C) was slightly higher than that reported by others (38.7 °C [9], 38.5 °C to 38.8 °C [10], 38.7 °C to 38.8 °C [3]). During the hot period mean RT was higher (39.4 ⫾ 0.6 °C) than the ones reported above for moderate conditions, which is very likely to be due to the greater impact of THI on BT
Table 2 Difference of the prevalence of fever defined by three different thresholds (ⱖ 39.5 °C, ⱖ 39.7 °C, and ⱖ 40.0 °C) measuring rectal (07.00 ⫾ 1 h and 17.00 ⫾ 1 h) and continuously (00.00 h to 24.00 h) measured vaginal temperature between the moderate (temperature humidity index: 59.8 ⫾ 3.8) and hot period (temperature humidity index: 74.1 ⫾ 4.4). Temperature
Rectal
Vaginal
Threshold (°C) ⱖ39.5 ⱖ39.7 ⱖ40 ⱖ39.5 ⱖ39.7 ⱖ40
Periods Moderate (%)
Hot (%)
P value
7.6 5.3 3.4 16.5 9.3 4.0
46.6 36.1 17.1 54.1 42.6 23.7
⬍0.001 ⬍0.001 ⬍0.001 ⬍0.001 ⬍0.001 ⬍0.001
Threshold (°C) ⱖ39.5 ⱖ39.7 ⱖ40 ⱖ39.5 ⱖ39.7 ⱖ40
Periods Moderate (%)
Hot (%)
P value
7.4 4.1 2.7 10.0 5.9 3.3
28.1 21.4 8.8 33.0 21.2 8.6
⬍0.001 ⬍0.001 0.02 ⬍0.001 ⬍0.001 ⬍0.001
under hot environmental conditions than under moderate ones [20]. This hypothesis is supported by the higher correlation of THI and RT and VT, respectively, during the hot period (RT: r ⫽ 0.46; VT: r ⫽ 0.40) compared with the moderate period (RT: r ⫽ 0.05; VT: r ⫽ 0.01). Furthermore, in the three studies stated above no systematic continuous measurement of the climatic condition (AT, RH, THI) prevailing on the farm was performed. To our knowledge this is the first study to monitor THI and quantify effects of hot weather conditions on physiological BT ranges. Mean VT during the moderate period in this study was higher (39.2 ⫾ 0.3 °C) than mean RT in this study (38.9 ⫾ 0.4 °C) and higher than the results from the studies cited above. Furthermore, mean VT during the hot period (39.6 ⫾ 0.4 °C) was higher than RT during the hot period (39.4 ⫾ 0.6 °C). This phenomena has been described before [23] and might be due to the fact that the logger is deeper in the body when inserted into the vagina than the probe of a thermometer inserted in the rectum. In this study we determined the prevalence of fever in two different approaches. In the first approach we averaged the two measures of RT and the 24 hourly measures of VT, respectively, to one daily RT and VT and described the percentage of measures exceeding three different thresholds. This approach is comparable to Wagner et al. [3] who measured RT every 4 h and determined the prevalence of fever (i.e., at least one measurement reaching the defined threshold during the first 10 DIM) of 66% (39.5 °C), 49% (39.7 °C), and 17% (40.0 °C). Interestingly these prevalence rates are much higher than the ones estimated in this study during the moderate period, but comparable with the ones estimated during the hot period (Table 2). However, the
O. Burfeind et al. / Theriogenology 78 (2012) 2031–2038
2037
Table 4 Mean rectal and vaginal temperature of healthy dairy cows during the first 10 days in milk (DIM) during a moderate (N ⫽ 17 cows; temperature humidity index: 59.8 ⫾ 3.8) and a hot period (N ⫽ 15 cows; temperature humidity index: 74.1 ⫾ 4.4). Period
Moderate
Hot
DIM
2 3 4 5 6 7 8 9 10 2 3 4 5 6 7 8 9 10
Rectal temperature (°C), mean (95% confidence interval) Morning (07.00 h)
Evening (17.00 h)
38.6 (37.9–39.2) 38.6 (38.1–39.2) 38.7 (38.1–39.4) 38.8 (38.1–39.6) 38.9 (38.2–39.5) 38.9 (37.9–40.0) 39.0 (38.0–40.0) 39.0 (37.7–40.2) 38.8 (38.2–39.5) 38.9 (38.0–39.6) 39.1 (38.3–39.8) 39.1 (38.1–40.0) 39.3 (38.2–40.4) 39.4 (38.4–40.4) 39.4 (38.3–40.6) 39.3 (38.5–40.1) 39.4 (38.6–40.2) 39.1 (38.2–40.0)
38.8 (38.1–39.5) 38.7 (38.0–39.4) 39.0 (37.9–40.1) 39.0 (38.2–39.8) 38.9 (38.1–39.7) 38.8 (38.1–39.6) 38.9 (37.8–40.0) 38.8 (38.3–39.3) 38.9 (38.3–39.6) 39.6 (38.3–40.8) 39.6 (38.1–41.1) 39.5 (38.3–40.7) 39.6 (38.5–40.8) 39.6 (38.4–40.7) 39.7 (38.5–40.9) 39.5 (38.6–40.4) 39.5 (38.3–40.7) 39.5 (38.5–40.5)
study of Wagner et al. [3] was conducted from fall to spring (September 2005 to April 2006) in North Dakota (USA), a region with continental climate. Therefore, the prevalence should be compared with the ones during the moderate period and not the hot period in our study. A possible explanation could be that health was determined by a clinical examination [3] and not by serum haptoglobin concentrations as in our study. Therefore, a possible misclassification of sick cows as healthy might have increased the prevalence of fever. In the second approach we determined the prevalence of fever for RT based on the measurement conducted in the morning (07.00 ⫾ 1 h) and for VT based on the average temperature from 07.00 h to 11.00 h. This approach was simulating a common routine of herdsmen who measure RT in the morning hours. Also this protocol was comparable to Wenz et al. [9] who reported a prevalence of fever (i.e., at least one measurement above 39.5 °C during the first 10 DIM) of 15% for multiparous cows. This is slightly higher than the prevalence in this study calculated for the moderate period, but lower than for the hot period (Table 3). The study was conducted in Colorado (USA), from May to July 2005 and 2006. The AT ranged from 7.2 °C to 25 °C, and most likely included moderate and hot periods. The THI, however, was not calculated. This might be a possible explanation for the estimated prevalence of fever lying in between the ones estimated in our study during the moderate and the hot period. Furthermore, in
Vaginal temperature (°C), mean (95% confidence interval) 38.9 (38.3–39.5) 38.9 (38.3–39.6) 39.1 (38.3–39.9) 39.2 (38.3–40.1) 39.2 (38.5–40.0) 39.2 (38.5–40.0) 39.3 (38.4–40.2) 39.2 (38.4–40.0) 39.2 (38.6–39.8) 39.5 (38.3–40.7) 39.5 (38.4–40.6) 39.5 (38.5–40.5) 39.6 (38.5–40.7) 39.7 (38.6–40.9) 39.7 (38.6–40.9) 39.6 (38.7–40.6) 39.6 (38.6–40.6) 39.5 (38.5–40.5)
the study conducted by Wenz et al. [9] cows were defined as healthy based on the production group (i.e., infectious disease, metabolic/digestive disease, no disease) to which they belonged. While health events were identified by farm personnel, metritis was diagnosed by farm and study personnel but with different definitions of metritis with respect to RT. Cows identified with metritis by study personnel did not necessarily have an RT ⬎ 39.4 °C, whereas this was a necessary requirement in the diagnosis of metritis by farm personnel. Therefore, 34% (50 of 145 cows) that were diagnosed with metritis by study personnel but not by farm personnel did not received any antibiotic treatment and were considered healthy. The authors speculated that these cows might have had less severe disease and therefore were not detected by farm personnel [9]. However, this fact clearly illustrates the need of a more specific parameter to define cows as healthy (i.e., serum haptoglobin concentration) in a research setting studying BT in healthy dairy cows. In this study RT and VT were influenced by several factors (DIM, period, time of day). For RT an additional interaction of DIM and time of day was present, whereas two additional interactions (DIM and period, time of day and period) influenced VT. Wenz et al. [9] identified similar factors influencing RT (month of calving, DIM, parity, disease group), including primiparous cows and diseased cows in their analyses. Our intention, however, was to describe RT and VT in
2038
O. Burfeind et al. / Theriogenology 78 (2012) 2031–2038
healthy cows, defined by a serum haptoglobin concentration ⱕ 1.1 g/L.
[7]
4.1. Conclusions This study demonstrates that BT 10 days postpartum is influenced by several factors (DIM, climate, time of day). Therefore, these factors have to be considered when interpreting BT measures to identify sick cows. Furthermore, the prevalence of fever considering different thresholds is higher during hot than moderate periods. However, even in a moderate period healthy cows can exhibit BT that is considered as fever. The data support that to minimize false positives, RT should be routinely taken in the morning. During moderate periods (THI 59.8 ⫾ 3.8), 7.4% of cows exhibited RT ⱖ 39.5 °C and during hot period (THI 74.1 ⫾ 4.4), 28.1% of cows exhibited RT ⱖ 39.5°C. This fact clearly illustrates that fever alone should not be considered the decision criterion whether a cow is allocated to an antibiotic treatment, although it is the most important one that is objectively measurable.
[8]
[9]
[10]
[11] [12]
[13]
[14]
Acknowledgments
[15]
The authors thank the owners, the herdswoman, and the staff of the farm. This project was funded in part by Pfizer Animal Health, Veterinary Medicine Research and Development, UK.
[16] [17]
[18]
References [1] Smith BI, Risco CA. Management of periparturient disorders in dairy cattle. Vet Clin North Am Food Anim Pract 2005;21:503– 21. [2] Benzaquen ME, Risco CA, Archbald LF, Melendez P, Thatcher MJ, Thatcher WW. Rectal temperature, calving-related factors, and the incidence of puerperal metritis in postpartum dairy cows. J Dairy Sci 2007;90:2804 –14. [3] Wagner SA, Schimeck DE, Chend FC. Body temperature and white blood cell count in postpartum dairy cows. The Bovine Practitioner 2008;42:18 –26. [4] Zhou C, Boucher JF, Dame KJ, Moreira M, Graham R, Nantel J, et al. Multilocation trial of ceftiofur for treatment of postpartum cows with fever. J Am Vet Med Assoc 2001;219:805– 8. [5] Sheldon IM, Lewis GS, LeBlanc S, Gilbert RO. Defining postpartum uterine disease in cattle. Theriogenology 2006;65:1516 –30. [6] Vickers LA, Burfeind O, von Keyserlingk MA, Veira DM, Weary DM, Heuwieser W. Technical note: comparison of rectal
[19]
[20]
[21]
[22]
[23]
and vaginal temperatures in lactating dairy cows. J Dairy Sci 2010;93:5246 –51. Burfeind O, von Keyserlingk MA, Weary DM, Veira DM, Heuwieser W. Short communication: repeatability of measures of rectal temperature in dairy cows. J Dairy Sci 2010;93:624 –7. Suthar V, Burfeind O, Bonk S, Voigtsberger R, Keane C, Heuwieser W. Factors associated with body temperature of healthy holstein dairy cows during the first 10 days in milk. J Dairy Res 2012;79:135– 42. Wenz JR, Moore DA, Kasimanickam R. Factors associated with the rectal temperature of holstein dairy cows during the first 10 days in milk. J Dairy Sci 2011;94:1864 –72. Kristula M, Smith AM, Simeone A. The use of daily postpartum rectal temperatures to select dairy cows for treatment with systemic antibiotics. The Bovine Practitioner 2001;35:117–25. Martin SW. The evaluation of tests. Can J Comp Med 1977;41: 19 –25. Huzzey JM, Duffield TF, LeBlanc SJ, Veira DM, Weary DM, von Keyserlingk MA. Short communication: haptoglobin as an early indicator of metritis. J Dairy Sci 2009;92:621–5. Drillich M, Voigt D, Forderung D, Heuwieser W. Treatment of acute puerperal metritis with flunixin meglumine in addition to antibiotic treatment. J Dairy Sci 2007;90:3758 – 63. Drillich M, Beetz O, Pfützner A, Sabin M, Sabin HJ, Kutzer P, et al. Evaluation of a systemic antibiotic treatment of toxic puerperal metritis in dairy cows. J Dairy Sci 2001;84:2010 –7. Kadzere CT, Murphy MR, Silanikove N, Maltz E. Heat stress in lactating dairy cows: a review. Livest Prod Sci 2002;77:59 –91. Hahn GL. Dynamic responses of cattle to thermal heat loads. J Anim Sci 1999;77 Suppl 2:10 –20. Silanikove N. Effects of heat stress on the welfare of extensively managed domestic ruminants. Livest Prod Sci 2000;67: 1–18. Igono MO, Bjotvedt G, Sanford-Crane HT. Environmental profile and critical temperature effects on milk production of holstein cows in desert climate. Int J Biometeorol 1992;36:77– 87. West JW, Mullinix BG, Bernard JK. Effects of hot, humid weather on milk temperature, dry matter intake, and milk yield of lactating dairy cows. J Dairy Sci 2003;86:232– 42. Kendall PE, Nielsen PP, Webster JR, Verkerk GA, Littlejohn RP, Matthews LR. The effects of providing shade to lactating dairy cows in a temperate climate. Livest Sci 2006;103:148 –57. Hellmann K, Radeloff I. International cooperation on harmonisation of technical requirements for registration of veterinary medicinal products. VICH, 2000. http://www.fda.gov/downloads/Animal Veterinary/GuidanceComplianceEnforcement/GuidanceforIndustry/ UCM052417.pdf. Accessed August 2012. Kendall PE, Tucker CB, Dalley DE, Clark DA, Webster JR. Milking frequency affects the circadian body temperature rhythm in dairy cows. Livest Sci 2008;117:130 – 8. Burfeind O, Suthar VS, Voigtsberger R, Bonk S, Heuwieser W. Validity of prepartum changes in vaginal and rectal temperature to predict calving in dairy cows. J Dairy Sci 2011;94:5053– 61.