- Email: [email protected]
Herd-level risk factors for infectious diseases in Swedish dairy calves aged 0–90 days
Preventive Veterinary Medicine 68 (2005) 123–143 www.elsevier.com/locate/prevetmed
Herd-level risk factors for infectious diseases in Swedish dairy calves aged 0–90 days G.K. Lundborga, E.C. Svenssona,*, P.A. Oltenacub a
Department of Animal Environment and Health, Faculty of Veterinary Medicine and Animal Science, Swedish University of Agricultural Sciences, P.O. Box 234, SE-532 23 Skara, Sweden b Department of Animal Science, Cornell University, Ithaca, NY 14853, USA
Received 19 February 2004; received in revised form 18 November 2004; accepted 30 November 2004
Abstract The effect of environmental factors and management routines on the risk of diarrhoea, respiratory disease and other infectious diseases was investigated in 3081 heifer calves 0–90 days old in 122 Swedish dairy herds. The farmers kept records on cases of diseases in their heifer calves and in addition, project veterinarians clinically examined all calves every 2–3 months. At each visit, the veterinarians also measured the ammonia concentration and relative air humidity in the housing facilities for the calves. The cleanliness of the animals and their environment was recorded as a measure of the hygienic status of the farm. The presence or absence of draught (i.e. wind velocity > 0.5 m/s) was recorded twice during the study period. The effect of these factors, as well as the placing of the calf pens, the nature of the pen walls, air volume per animal, management factors (such as the status of the caretaker and feeding routines) and presence or absence of a bovine viral diarrhoea virus (BVDV) infection in the herd, was evaluated by means of a two-level variance component logistic model. The placing of calf pens along an outer wall was significantly associated with the risk of diarrhoea (odds ratio (OR): 1.92, P < 0.01). The risk for respiratory disease was significantly associated with an ammonia concentration below 6 ppm (OR: 0.42, P < 0.05) while the odds ratio for moderately to severely increased respiratory sounds was significantly associated with a BVDV infection in the herd (OR: 2.39, P < 0.05) and draught (OR: 3.7, P < 0.02). Absence of
* Corresponding author. Tel.: +46 511 67205; fax: +46 511 67204. E-mail address: [email protected] (E.C. Svensson). 0167-5877/$ – see front matter # 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.prevetmed.2004.11.014
124
G.K. Lundborg et al. / Preventive Veterinary Medicine 68 (2005) 123–143
draught was significantly associated with the risk for infectious diseases other than diarrhoea and respiratory disease (OR: 0.42, P < 0.01). # 2004 Elsevier B.V. All rights reserved. Keywords: Ambient environment; Bovine viral diarrhoea virus (BVDV); Dairy calf; Diarrhoea; Herd-level morbidity; Management; Respiratory disease; Risk factors
1. Introduction In the countries within the European Union, the interest of the general public in animal welfare has increased during the past decades (Kirkwood and Hubrecht, 2001). For the consumer, the conditions the animals are kept in are an increasingly important aspect of the quality of dairy products, making this a valuable sales argument for the dairy industry. To many consumers, the welfare of the dairy calf is as important as is that of the adult animal. However, for the farmer, the calf often comes second to the lactating cow. Suboptimal management of and a suboptimal environment for dairy calves are not only an animal welfare issue, but also have a direct economic impact on the farming enterprise as they are likely to result in a higher calf morbidity and lower growth rates. Gunn and Stott (1997) estimated the average loss per calf at risk in a herd with an outbreak of enteritis, accounting for calf mortality, loss in calf value, expenses for extra work, possible veterinary treatment and extra feed to compensate for the retained growth, to be approximately £33. They found the corresponding loss per calf at risk in a herd with an outbreak of pneumonia to be approximately £21 (Gunn and Stott, 1997). Andrews (2000) calculated the average extra rearing cost to be £43 for a diseased calf in a herd with an outbreak of pneumonia. Although management and the environment are known to play an important role in the clinical outcome of the multi-factorial infectious diseases that are the main health disorders in calves, knowledge about the effects of individual factors is still limited. Pneumonia has been associated with poor climatic conditions (Kilburn, 1967; Harry, 1978; Kiorpes et al., 1988). However, few studies have objectively measured air quality parameters such as ammonia concentration and relative humidity and examined their effect on the risk of respiratory disease in calves. Similarly, draught is often mentioned as a risk factor for calf morbidity, but we have not come across any study that more objectively examined its effect. The literature offers conflicting results about the effects on calf health of housing calves separately (Simensen and Norheim, 1983; Fourichon et al., 1997; Virtala et al., 1999) and of the sex and status of the calf caretaker (Oxender et al., 1973; Speicher and Hepp, 1973; Hartman et al., 1974; Jenney et al., 1981; Hagstad et al., 1984; James et al., 1984; WaltnerToews et al., 1986a), and no firm conclusions have been drawn about the relationship between calf health and herd size (Oxender et al., 1973; Speicher and Hepp, 1973; Hartman et al., 1974; James et al., 1984; Waltner-Toews et al., 1986a; Curtis et al., 1988). With regard to the effects of management and feeding procedures, the literature is surprisingly sparse. Introduction of bovine viral diarrhoea virus (BVDV) into the dairy herd has been associated with impaired calf health (Barber et al., 1985; Houe and Meyling, 1991; Larsson
G.K. Lundborg et al. / Preventive Veterinary Medicine 68 (2005) 123–143
125
et al., 1994; Moerman et al., 1994) and de Verdier Klingenberg et al. (1999) demonstrated a decreased incidence of calf diarrhoea after strict closure and eradication of BVDV infection in a dairy herd. However, the reports mentioned above all cover the situation in a single or a few herds only and there are few large-scale studies on the effect of BVDV status on calf morbidity. A large-scale study performed by Ersbøll et al. (2001) reported an increased calf mortality rate in herds with a BVDV infection. The aim of the present paper was to describe the herd-level morbidity and evaluate the effect of some potential herd-level risk factors on the incidence of infectious diseases in 0– 90-day-old dairy calves reared under Swedish conditions.
2. Material and methods 2.1. Animals The study was performed in all heifer calves born on 122 farms in the county of Skaraborg in south-western Sweden during 1998, in total 3081 calves. All the farms were enrolled in the official milk recording programme. All farmers had a herd size of between 28 and 94 cows (median 47.5 cows), housed their young calves in individual or group pens and were considered by the large animal practitioners and farm advisors in the region to be capable of keeping thorough records of their calves’ performance. A more detailed description of the selection of the participating farms can be found in a previous paper (Svensson et al., 2003). After a statistical comparison between the 122 farms and a random sample of 877 Swedish dairy herds of similar size (Pettersson et al., 2001; Svensson et al., 2003), the farms were considered a reasonably representative sample of Swedish dairy farms with 28–94 cows. They were mostly run exclusively by family members, but 28% had one or two employees. The population of heifer calves monitored was dynamic throughout the study period. Calves entered the monitoring phase of the study at birth and left at 90 days of age. Altogether, 108 (3.5%) calves died or were slaughtered and 26 (0.8%) were sold or lost to the study before 90 days of age. Calves were generally kept indoors and housed in single pens, small group pens for between 3 and 8 calves, with manual feeding of the milk, or in large group pens for 6–30 calves with an automatic milk-feeding system. Sixty-three percent of the calves were removed from their mothers immediately after birth; the rest were allowed to stay with their mothers for 8 h–4 days. Almost all calves received colostrum from their own mother (98%). Pooling of colostrum is very rarely used in Sweden. Colostrum exclusively from the first milking was used for the calves’ first two meals by 36% of the farmers. Calves received whole milk until weaning in 45% of the herds and milk replacements in 45% of the herds. In the remaining herds, they were fed a combination of milk replacement and whole milk. The milk volume given ranged from 3 to 8 L (median 5.0 L) per day. The calves were weaned at a median of 9.0 weeks of age (20–80th percentiles: 8–11 weeks). They were given free access to hay. Concentrates, mainly as pelleted calf feed or as crushed grain solely or in combination with protein feed, were given starting from day 1 to day 112 (median day 7). The concentrates were given ad libitum in 88% of the herds.
126
G.K. Lundborg et al. / Preventive Veterinary Medicine 68 (2005) 123–143
2.2. Data collection The farmers were requested to keep records on cases of disease (treated as well as untreated) in their heifer calves on special individual health cards. On the health cards, they also continuously recorded for each calf date of birth, place of birth (BIRTHPLACE; individual maternity pen, tie stall, cubicle or group maternity pen, or pasture), time of birth (BIRTHTIME; day or night), breed (BREED), method of feeding colostrum (COLFEED; suckling or by bucket/nipple feeder), origin of colostrum (COLORIGIN; a primiparous cow, a second calving cow, a cow that has calved three or more times), hours from birth to first colostrum feeding (COLTIME), housing system used for the calves (HOUSING; individual pen, small-group pen with manual feeding of the milk or large-group pen with an automatic milk feeding system) and supervision of calving (SUPERVIS; yes or no). For a more detailed description and for the distributions of these calf-level variables, see Svensson et al. (2003). Based on our definitions of respiratory disease and diarrhoea, the farmers recorded these two diseases directly on the individual health cards. ‘‘Diarrhoea’’ was defined as a faecal consistency that for 2 days or more was softer than is normally observed in cattle of the age in question. ‘‘Respiratory disease’’ was defined as coughing or sneezing for more than 2 days, as severely increased respiratory sounds at lung auscultation or as moderately increased respiratory sounds together with additional signs, such as coughing or nasal discharge. For all other disease conditions, the farmers recorded the symptoms and the project veterinarians on their visits (every second month) determined the diagnosis, based on mutual definitions. On these visits, the project veterinarians also clinically examined all calves, auscultated their lungs, checked the notes kept by the farmers and recorded prevalent diseases not detected by the farmers in the health records. Recorded infectious diseases were divided by the veterinarians into the following categories: arthritis, omphalophlebitis/umbilical abscess, weak calf syndrome and ‘‘other infectious disease’’ (e.g. abscesses and ringworm). ‘‘Arthritis’’ was defined as swelling of one or more joints accompanied by lameness and fever. ‘‘Omphalophlebitis/umbilical abscess’’ was defined as a warm swelling or abscess formation associated with the navel. Weak calf syndrome was diagnosed if an animal had inappetence and dullness during at least 2 days, with or without an increased body temperature, but lacked other symptoms of disease. Lungauscultation findings were ranked as normal respiratory sounds or as mildly, moderately or severely increased respiratory sounds. At each farm visit, the veterinarians also measured the ammonia concentration (AMMONIA) and the relative air humidity (HUMIDITY) in the housing facilities for the calves. The same measure points were used at all visits: at the nose level of standing animals in the most central pen used for calves 0–90 days of age. If there were several types of calf pens in the herd (e.g. calves housed in group pens were most often kept in single pens during their first weeks of life and singled-housed calves were sometimes moved to group pens before 90 days of age), one measure point for each pen type was used. Recordings were made on visits from January 1998 to April 2000, so that in total 8–12 measurements were made for each pen type and herd, reflecting conditions during different months of the year. The ammonia concentration was measured using a mobile Dra¨ ger1 gas monitor Pac III (Dra¨ gerwerk AG, Lu¨ beck, Germany), while relative humidity was
G.K. Lundborg et al. / Preventive Veterinary Medicine 68 (2005) 123–143
127
measured using a Testo1 term hand instrument (Testoterm GmbH & Co., Lenzkirch/ Schwarzwald, Germany). The cleanliness of the animals (ANIMCLEAN) and of the bedding (BEDCLEAN), the concentrate and water buckets (FBUCKCLEAN and WBUCKCLEAN) and the pen walls (PENCLEAN) were recorded at each visit as measures of the hygienic status of the farm. On each of the first five visits to a farm, the two next youngest animals were chosen until 10 individuals per farm had been selected. During each farm visit, the project veterinarian monitored the cleanliness of each of these 10 animals as well as of the walls and bedding of their present pen. Their concentrate and water buckets were also examined and their cleanliness scored. The animals were given one cleanliness score for each of eight different body parts (front claws and hind claws, front legs below the carpus, hind legs below the hock, brisket, thighs, belly and tail). The scores ranged from 0 to 100% and represented the percentage of the body part that was contaminated with faeces. A value of 0% described a clean area, 5–30% a mildly contaminated, >30–70% a moderately contaminated and >70% a heavily contaminated body area. The cleanliness of pen walls, bedding, and the concentrate and water buckets were scored similarly from 0 to 100%. Before the start of the study and twice during the study, simultaneous monitoring was made by the three project veterinarians in order to standardise their scales. Once in the summertime and once in the wintertime, presence or absence of draught (DRAUGHT) was determined in each of the calf pens on the farms using a smoke bottle (Titranetetrachloride RFA; AB Regin, Sweden). Draught was considered present (score 1) if wind velocity exceeding 0.5 m/s was detected in any of the pens. The project veterinarians also measured the air volume in the calf premises and calculated the number and category of animals per building and the number of animals per pen. They also recorded information on placing of the calf pens within the building and nature of pen walls (e.g. possibilities for direct nose contact between calves from different pens). Based on their experiences during the study, the three project veterinarians jointly graded the capability of the farmers (REGCAPACITY) to keep accurate records of diseases in their calves as ‘‘less good’’, ‘‘good’’ or ‘‘excellent’’. Data on feeding and management routines were collected by the project veterinarians by means of interviews with the farmers at farm visits. The farmers were asked about the routines for cleaning the milk buckets (BUCKRINSE), the storage of the milk buckets between meals (BUCKSTORE), the use of the milk buckets for the calves (BUCKUSE), the routines for cleaning the concentrate buckets (CONBUCKCLEAN), whether concentrate not eaten was removed (CONCREMOVED), the routines for dosing the milk powder (MILKDOSE), milk type used (MILKTYPE), whether they gave the cows ADE-vitamins before calving (VITAMIN), whether they heated the whole milk by adding hot water (WATERMILK) and weaning routines (WEANHOW). For a description of the categories within each variable, see Table 1a–c. The farmers were also interviewed about feeding routines and routines at calving and weaning, as well as about the status of the caretakers and the number of animals bought annually (see Table 1a–c). Data on the BVDV status of the herds were collected from the Swedish BVDV control programme. According to the control programme, a herd is certified as BVDV-free if two samples (i.e. samples of bulk milk, pooled milk samples from primiparous cows or individual blood samples) collected with a 7 months’ interval are seronegative for the virus.
128 Table 1 (a) Management variables, (b) feeding hygiene-related variables and (c) feeding-related variables constructed from information gathered through interviews with farmers, and (d) variables constructed from observations or measurements made by project veterinarians during visits on 122 dairy farms in south-western Sweden
(a) Management variables BUYANIM Number of animals bought annually
CALVINGPEN Percentage of times when the calving pen was cleaned between consecutive calvings CARETAKER Number of persons involved in the care of the calves
CHORE Percentage of cases when the feeding of the calves was done at the same time as another chore COWCALF Proportion of the calves that were held with the dam sometime after birth COWNO Number of cows in the herd
Categories
Number of herds
Included in multivariate analysesa
1 = 5 2 = 1 to <5 3=0
11 13 98
I
1 = >75 2 = >25 to <75 3 = 25
39 12 8
R, A, I
1 = 3 2=2 3=1
21 60 41
D, I
1 = >75 2 = >25 to <75 3 = 25
20 16 86
R, A
1=0 2 = >0 to <1 3=1
60 44 18
D, R, A, I
1 = 59 2 = 40 to <59 3 = <40
31 64 27
D, R, A, I
G.K. Lundborg et al. / Preventive Veterinary Medicine 68 (2005) 123–143
Variable
EMPLOYEE Percentage of calf caretaking performed by an employee
VITAMIN Whether cows were routinely supplied with ADE vitamins before calving or not WEANAGE Age at weaning (weeks)
WEANHOW Routines for weaning
(b) Feeding hygiene-related variables BUCKRINSE Routines for rinsing milk buckets
BUCKSTORE Routines for storing milk buckets between the meals
11 17 94
1 = >75 2 = >25 to <75 3 = 25
32 30 60
D, R, A, I
1 = yes 2 = no
23 99
D, R, A, I
1 = 10 2 = >8 to <10 3 = 8
39 29 54
1 = weaning abruptly 2 = weaning by reducing milk meals to one per day 3 = weaning by offering less milk per meal 4 = weaning by adding increasing amounts of water to the milk
28 27
1 = buckets rinsed after every meal 2 = buckets rinsed less often than after every meal
100 20
43 24
22 48 49
D 129
1 = combination 2 = stored separately 3 = stacked when stored
D, R, A
G.K. Lundborg et al. / Preventive Veterinary Medicine 68 (2005) 123–143
FEMALE Percentage of calf caretaking performed by a woman
1 = >75 2 = >25 to <75 3 = 25
130
Table 1 (continued ) Variable
CONBUCKCLEAN Routines for cleaning concentrate buckets
CONCREMOVED Whether concentrates not eaten are removed or not (c) Feeding-related variables COLFIRSTMILK Percentage of cases in which colostrum from the first milking was used for more feeds than the first COLLITRES Litres of colostrum fed on the first day
Number of herds
1 = one unique bucket per calf 2 = one bucket per calf and meal 3 = one bucket for several calves during the same meal
15 30 72
1 = buckets cleaned every second week 2 = buckets cleaned once a month 3 = buckets cleaned when calf leaves the pen 4 = buckets cleaned less often than every 7th week
17
Included in multivariate analysesa
29 27 31
1 = yes 2 = no
92 17
1 = >75 2 = >25 to <75 3 = 25
53 17 52
D, R, A, I
1 = 8 2 = >4 to <8 3 = 4
11 75 36
D, R, A, I
G.K. Lundborg et al. / Preventive Veterinary Medicine 68 (2005) 123–143
BUCKUSE Routines for use of milk buckets
Categories
CONCAGE Age at which the calves were first offered concentrates (weeks)
MILKDOSE Dosage of milk powder in milk replacement
MILKTYPE Type of milk
TEMPCHECK Percentage of meals at which the temperature of the milk was checked WATERMILK Whether the whole milk was heated by adding hot water or not WATERAGE Age at which the calves were first offered water (weeks)
15 34 73
D
1 = >7 2 = >1 to 7 3 = 1
11 49 60
D
1 = according to manufacturer’s instructions 2 = different from manufacturer’s instructions
43 15
1 = combination 2 = whole milk 3 = milk replacement
9 56 56
1 = >75 2 = >25 to <75 3 = 25
75 12 28
D
1 = yes 2 = no
14 40
R, A
1 = >4 2 = 2 to <4 3 = <2
47 28 47
G.K. Lundborg et al. / Preventive Veterinary Medicine 68 (2005) 123–143
HAYAGE Age at which the calves were first offered hay (days)
1 = >2 2 = >1 to 2 3 = 1
(d) Variables constructed from observations or measurements made by project veterinarians 131
132
Table 1 (continued ) Variable
ANIMCLEAN Cleanliness of the animals (percentage of the animal surface contaminated with faeces) AMMONIA Ammonia concentration (ppm) BEDCLEAN Cleanliness of the bedding (percentage of the bedding contaminated with faeces) BUILDING Housing of calves in relation to older cattle
BVDV BVDV-statusb
CONTACT Possibilities for nose contact with calves from another pen
1 = >20 2 = 20
Number of herds 60
Included in multivariate analysesa R, A
61 1 = >30 2 = 30
8 114
1 = 6 2 = <6
38 64
1 = >30 2 = 30
28 94
1 = calves housed in the same building section as cows and/or older young stock 2 = calves housed separately 1 = BVDV introduced into the herd during the study period 2 = BVDV present in the herd at the start of the study period 3 = BVDV not present in the herd 1 = no 2 = yes
R, A, I
103
19 12 32 78 104 17
D, R, A, I
G.K. Lundborg et al. / Preventive Veterinary Medicine 68 (2005) 123–143
AIRVOLUME Air volume (m3 air/animal unit) (calves = 0.2 units; young stock >6 months = 0.5; cow or heifer near calving = 1.0 units)
Categories
DRAUGHT Draught (defined as a wind velocity exceeding 0.5 m/s detected in any pen)
HUMIDITY Relative air humidity (%) PENCLEAN Cleanliness of the pen (%) PENPLACE Placing of the calf pens
REGCAPACITY Farmers’ capability of keeping accurate records
WBUCKCLEAN Cleanliness of the water bucket (percentage of buckets that were considered clean) a b
17
D, R, A, I
53
1 = >30 2 = 30
47 64
1 = >80 2 = 80
13 109
1 = >30 2 = 30
55 67
D, R, A, I
1 = against an outer wall 2 = separated from the outer walls 3 = combinations
29 42 50
D, R, A, I
1 = less good 2 = good 3 = excellent
20 82 20
D, R, A, I
1 = >30 2 = 30
8 43
D, I
52
D, I
R, A, I
Variables examined for their effects on risk of diarrhoea (D), respiratory disease (R), increased respiratory sounds (A) and other infectious disease (I). Information retrieved from the Swedish official BVDV (bovine viral diarrhoea virus) control programme.
G.K. Lundborg et al. / Preventive Veterinary Medicine 68 (2005) 123–143
FBUCKCLEAN Cleanliness of the concentrate buckets (percentage of the buckets that were considered clean)
1 = 1 (draught recorded both in summer and winter) 2 = >0 to <1 (draught recorded in summer or winter) 3 = 0 (no draught recorded)
133
134
G.K. Lundborg et al. / Preventive Veterinary Medicine 68 (2005) 123–143
For a closer description of the eradication programme, see Olsson et al. (1993) and Alenius et al. (1997). 2.3. Editing of data and statistical analysis In cases where results from more than one lung-auscultation were recorded in a calf prior to 91 days of age, the last recording was used. No findings recorded in calves younger than 31 days were used in the study. The findings at lung auscultation were dichotomised into normal to mildly increased respiratory sounds (0) and moderately to severely increased respiratory sounds (1). A variable SEASON was constructed based on the date of birth. Birth dates during the period from 1 May to 31 August was defined as summer dates, 1 September to 30 November defined as autumn dates and 1 December to 30 April as winter dates. Herd AMMONIA and HUMIDITY values were constructed from mean values of all recordings in the herd. If more than one pen type was used for calves in the herd, a mean value was first calculated for each pen type. An overall weighted mean was thereafter calculated based on the percentage of weeks from day 0 to day 90 that the calves usually spent in each pen type. Herd DRAUGHT scores were calculated as the mean of the summer and winter scores in each herd (scores between 0 and 1). Herd-level scores were also constructed for the five cleanliness parameters. For each of the 10 individuals monitored per farm, median body part scores were first constructed based on all scores recorded for the particular body part. Thereafter, the median body part scores were each used to construct median animal scores, calculated as the median score of the three body parts with the highest scores (if hind and front claws were among these three body parts, only the highest claw score was included). An ANIMCLEAN score for the herd was calculated as the median of all animal scores in the herd. PENCLEAN, BEDCLEAN, FBUCKCLEAN and WBUCKCLEAN scores for each of the herds were calculated similarly. Continuous predictor variables were transformed to categorical variables; cut-off points were based on biological knowledge, practical common routines or quartiles. The categories and their distributions can be seen in Table 1a–d. The herd-level variables were initially screened in univariate analyses for each of the four outcome variables: diarrhoea, respiratory disease, ‘‘other infectious disease’’ and moderately to severely increased respiratory sounds, using the procedure PROC LOGISTIC in SAS (SAS Institute Inc., 2000). All herd-level variables that were associated with an outcome variable at P 0.2 using the F-test were then transferred to the multi-level statistical software MLwiN (MLwiN, 2002; Rasbash et al., 2002). In addition, we considered calf-level factors found to significantly affect the incidence risk of the outcome variables in a previous study based on the same material (Svensson et al., 2003): BIRTHPLACE, BREED, BIRTHTIME, COLFEED, COLORIGIN, COLTIME, HOUSING, SEASON and SUPERVIS. As with herd-level variables, calf-level variables associated with the outcome at P 0.2 in a univariate analysis using the F-test were transferred to MLwiN. For each binary outcome, a two-level (calf; herd) variance components logistic model was built, using second-order penalised quasi-likelihood estimations (Goldstein and
G.K. Lundborg et al. / Preventive Veterinary Medicine 68 (2005) 123–143
135
Rasbach, 1996) with the restricted iterative generalised least-squares method (Goldstein, 1995). The models were defined by the following equations: Yij ¼ pij þ eij logitðpij Þ ¼ b0j þ
X
bm Xmij
b0j ¼ b0 þ uj where Yij is the disease status of calf i in herd j, pij the probability that calf i in herd j was diseased, eij a random term at the calf-level, b0j the intercept of herd j, bm the regression coefficients expressing effects of the included predictors Xmij, b0 the mean intercept among all herds, and uj is the random term at the herd-level. The final models were built by a manual forward stepwise procedure, including one factor at a time and excluding factors that were not significant, according to the approximate two-sided Wald (large-sample joint chi-squared) test (P 0.05) before proceeding. Finally, biologically relevant first-order interactions were tested one at a time. No significant (Wald P 0.05) interactions were found. The selection of variables to be included was partly based on subjective judgement, favouring model simplicity and taking confounding factors into account; thus, in spite of the increased risk for spurious associations, a few cases of borderline significant effects (0.05 < P 0.10) were included if they were found to act as confounders and stabilise the model upon inclusion. A binomial distribution was assumed, constraining the residual variance to 1, but the degree of overdispersion was checked by re-running the models without the constraint. Level 2 residuals were checked graphically. Odds ratios (ORs) were calculated from the final parameter estimates. The proportion of unexplained variation residing at the herd-level was calculated as herd variance/(herd variance + p2/3), according to the latent-variable method (Goldstein et al., 2002). The total numbers of observations at herd as well as at calf level are given in Table 2a–d.
3. Results 3.1. Herd-level morbidity The median total herd disease incidence risk was 21.6%. However, the incidence risk differed considerably between the herds and ranged from 0.0 to 57.6%, the central range (20–80th percentiles) being 10.2–34.1%. In two herds, no morbidity was recorded; in 22 herds, no diarrhoea was recorded and in 55 herds, no cases of respiratory disease were recorded. The corresponding numbers for arthritis, other infectious disease and omphalophlebitis were 106, 87 and 93 herds, respectively. The herd disease incidence risks for the most common diseases are presented in Table 3. 3.2. Risk factors The risk for calves of suffering from diarrhoea was significantly (P < 0.01) associated with the placing of the calf pens against an outer wall, compared to calves in herds in which
136
G.K. Lundborg et al. / Preventive Veterinary Medicine 68 (2005) 123–143
Table 2 Effects of herd-level factors on incidence of (a) diarrhoea, (b) respiratory disease, (c) moderately to severely increased respiratory sounds, (d) infectious diseases other than diarrhoea and respiratory disease, from birth to 90 days of age in calves in 115–121 dairy herds in south-western Sweden from January 1998 to March 1999 Variablea
Levels
(a) Diarrhoea Intercept b,c
b
S.E. (b)
OR
95% CI (OR)
P
2.3
0.42
–
–
–
BREED
SRB SLB SLB SRB Cross-breed beef
0 0.47 0.25 0.88
– 0.17 0.48 0.78
– 0.63 0.78 0.41
– 0.45–0.87 0.30–2.00 0.09–1.91
<0.05
COLORIGINb
First-lactation cow Second-lactation cow Third or more lactation cow
0 0.53 0.31
– 0.18 0.15
– 0.59 0.73
– 0.41–0.84 0.55–0.98
<0.01
PENPLACE
Against the outer wall Separated from the outer wall Combinations
0 0.65 0.62
– 0.24 0.23
– 0.52 0.54
– 0.33–0.84 0.34–0.84
<0.01
REGCAPACITY
Less good Good Excellent
0 0.26 0.64
– 0.26 0.31
– 1.30 1.90
– 0.78–2.16 1.03–3.48
<0.10
TEMPCHECK
>25 to <75% of meals >75% of meals 25% of meals
0 0.39 0.70
– 0.37 0.33
– 1.48 2.01
– 0.72–3.05 1.05–3.85
<0.10
0.51
–
–
–
0 0.87
– 0.37
– 2.39
– 1.16–4.93
<0.05
(b) Respiratory disease Intercept
3.0
AMMONIA
6 ppm <6 ppm
HOUSINGd
Large-group pen Single pen Small-group pen
0 0.84 0.87
– 0.33 0.41
– 0.43 0.42
– 0.23–0.82 0.19–0.93
<0.05
REGCAPACITY
Less good Good Excellent
0 0.93 0.89
– 0.39 0.52
– 0.39 0.41
– 0.18–0.85 0.15–1.14
<0.10
SEASONd
Summer Autumn Winter
0 0.90 0.66
– 0.23 0.22
– 2.46 1.94
– 1.57–3.86 1.26–2.98
<0.001
(c) Moderately to severely increased respiratory sounds Intercept
0.15
0.80
–
–
–
0 0.87 0.03
– 0.36 0.59
– 2.39 1.03
– 1.16–4.93 0.32–3.26
<0.05
Detected in summer and winter Detected in summer or winter Not detected
0 0.41 1.30
– 0.46 0.50
– 0.66 0.27
– 0.27–1.63 0.10–0.73
<0.02
Large-group pen Single pen Small-group pen
0 1.70 2.00
– 0.38 0.51
– 0.18 0.14
– 0.08–0.37 0.05–0.37
<0.001
BVDV
Not present in the herd Present in the herd at start Introduced during study
DRAUGHT
HOUSINGe
G.K. Lundborg et al. / Preventive Veterinary Medicine 68 (2005) 123–143
137
Table 2 (Continued ) Variablea
Levels
b
S.E. (b)
OR
95% CI (OR)
P
REGCAPACITY
Less good Good Excellent
0 0.89 1.50
– 0.40 0.62
– 0.41 0.22
– 0.19–0.90 0.07–0.75
<0.05
WEANHOW
Weaning Weaning Weaning Weaning
0 0.69 1.10 1.10
– 0.48 0.55 0.49
– 0.50 0.33 0.33
– 0.20–1.29 0.11–0.98 0.13–0.87
<0.10
–
–
–
by adding more water squarely by reducing no of meals by reducing volume/meal
(d) Infectious diseases other than diarrhoea and respiratory disease Intercept 4.0 0.44 f
BIRTHTIME
Day Night
0 0.42
– 0.21
– 1.52
– 1.01–2.30
<0.05
DRAUGHT
Detected in summer and winter Detected in summer or winter Not detected
0 0.15 0.87
– 0.40 0.39
– 1.16 2.39
– 0.53–2.54 1.11–5.13
<0.01
COLORIGINf
Third or more lactation cow Second-lactation cow First-lactation cow
0 0.51 0.08
– 0.25 0.25
– 1.67 1.08
– 1.02–2.72 0.66–1.76
<0.10
PENPLACE
Against the outer wall Separated from the outer wall Combinations
0 0.69 0.09
– 0.31 0.29
– 0.50 0.92
– 0.27–0.92 0.52–1.62
<0.10
a
Abbreviations found in Table 1a–d. Extra explanatory calf-level variable, n = 2849. c SRB: Swedish Red and White breed; SLB: Swedish Holstein; SLB SRB: all cross-breeds between Swedish Red and Whites and Swedish Holsteins; Cross-breed beef: cross-breeds with a dam of Swedish Red and Whites or Swedish Holsteins and a sire of beef breed. d Extra explanatory calf-level variable, n = 3037. e Extra explanatory calf-level variable, n = 2049. f Extra explanatory calf-level variable, n = 2923. b
calf pens were separated from an outer wall (OR: 1.92) (Table 2a). A decreased risk for calves of suffering from respiratory disease was found to be significantly (P < 0.05) associated with a higher ammonia level (OR: 0.42) (Table 2b). The OR for moderately to severely increased respiratory sounds was significantly associated with BVDV, DRAUGHT and REGCAPACITY (Table 2c). A BVDV infection present in the herd at the start of the study increased the risk for the calves of suffering from moderately to severely increased respiratory sounds (OR: 2.39, P < 0.05) compared to calves in BVDVfree herds. The presence of draught also increased the risk for the calves of suffering from moderately to severely increased respiratory sounds (OR: 3.7, P < 0.02) compared to calves in herds in which draught was not detected. Calves in herds in which the farmer had an excellent record-keeping capacity had a decreased risk of suffering from moderately to severely increased respiratory sounds (OR: 0.22, P < 0.05). The absence of draught was significantly associated with an increased risk for the calves of suffering from an infectious disease other than diarrhoea and respiratory disease (OR: 2.39, P < 0.01) (Table 2d).
138
G.K. Lundborg et al. / Preventive Veterinary Medicine 68 (2005) 123–143
Table 3 Herd incidence risks (% calves/herd) of the most common diseases in 0–90-day-old calves from 122 dairy herds in south-western Sweden from January 1998 to March 1999
Arthritis Diarrhoea Increased respiratory sounds Omphalophlebitis Respiratory disease Other infectious diseases
Median (%)
Range (%)
20–80th percentile (%)
0.0 7.8 14.3 0.0 3.0 0.0
0.0–5.6 0.0–39.4 0.0–72.4 0.0–17.6 0.0–51.7 0.0–36.4
0.0–0.0 2.3–17.6 5.6–29.7 0.0–3.0 0.0–11.4 0.0–4.3
There was a tendency (P < 0.10) for the risk of diarrhoea to be increased for calves in herds in which the temperature of the milk was checked at less than 20% of the feeding times compared with calves in herds where the temperature check was performed 20–70% of the occasions. There was a tendency (P < 0.10) that the risk of finding moderately to severely increased respiratory sounds was greater in herds in which the weaning practice was to add increasing amounts of water to the milk compared with herds that weaned their calves by serving less and less milk per meal. We did not find an association between diarrhoea and method of weaning. In the present study, no significant association was found between the cleanliness of the bedding and calf morbidity.
4. Discussion 4.1. Herd-level morbidity The total herd-level morbidity found in the present study was of a similar level as that previously reported by Waltner-Toews et al. (1986b) and, similarly to that study, most herds had a fairly low disease incidence risk while a few herds had a high incidence risk. The skewed distribution may partly be due to the contagious nature of infectious diseases, but may also reflect differences in management and the skills of the herdsman. 4.2. Risk factors In several studies, high ammonia levels have been suggested to increase the risk of respiratory disease through irritation of the respiratory epithelium and decreasing the mucociliary flow (Kilburn, 1967; Harry, 1978; Kiorpes et al., 1988). Our data suggest a reversed relationship, which is difficult to explain. The ammonia levels recorded seem reasonable in relation to other ambient environment factors measured. However, the highest ammonia concentrations were found in pens with slatted floors, to which calves were moved only shortly before 90 days of age. For ammonia to have an effect on the level of respiratory disease, a longer exposure time may be needed, extending beyond the study period. The fact that the herd means included measurements in pens used only during the
G.K. Lundborg et al. / Preventive Veterinary Medicine 68 (2005) 123–143
139
last part of the study period may therefore have obscured the results. Our examination of the data shows that this cannot explain the unexpected result. However, the observed effect of ammonia may have been confounded by some unknown and unrecorded factor. The effects of season and caretaker were included in the model. AMMONIA showed a fairly small variation (range: 0.3–13.5 ppm; 20–80th percentiles: 3.3–6.7 ppm), which may have affected the results. Martig et al. (1976) recommended a maximum of 25 ppm ammonia for calves while Wathes (1998) recommended a 20 ppm limit for pigs. Both values far exceed the allowed level, according to the Swedish animal welfare legislation, which is 10 ppm. Another finding which is equally difficult to explain is the significant association between the detection of draught and the decreased risk of other infectious disease in a herd. Calves in herds where draught was detected were found to suffer an increased (OR: 3.7) risk of having moderately to severely increased respiratory sounds at lung auscultation, compared with calves in herds without draught. Wathes et al. (1983) defined draught as an air speed faster than 0.3 m/s, which leads to enhanced heat losses and chilling in cold weather. Martig et al. (1976) suggested that for the well being of the animals, the air speed should not exceed 0.2 m/s. The stress of heat loss and chilling has been found to negatively affect the calf’s immune system, leading to a higher susceptibility of different pathogens (Jennings and Glover, 1952). In our study in herds whose calf pens were placed along an outer wall, the calves had an increased risk of suffering from diarrhoea (OR: 1.92) and tended also to have an increased risk of other infectious disease (P < 0.10). The association between the placing of the calf pens and disease could not be explained by draught. Alternative explanations for the effect may be related to cold radiation, high relative humidity and damp bedding conditions. Calves in herds where antibodies against bovine virus diarrhoea virus or persistently infected animals were present at the start of the study ran an increased risk of having moderately to severely increased respiratory sounds at lung auscultation compared to herds free of the infection. Larsson et al. (1994) found calf mortality and the percentage of calves treated by veterinarian for respiratory disease and/or enteritis to be significantly higher in calves born during the period of BVDV introduction to the herd than in calves born the previous year or the year after (when calves persistently infected with the virus was present for most of the period). Similar findings had previously been reported by Barber et al. (1985). The primary infection with BVDV has been shown to alter the function of neutrophils and depress the proliferative rate of lymphocytes when tested in vitro (Roth and Kaeberle, 1983). Tra˚ ve´ n et al. (1991) found a transient leukopenia and lymphopenia in the acute phase of the infection. Therefore, the virus appears to be an immunosuppressive agent which may increase the susceptibility of the host to other respiratory or enteric pathogens (Potgieter et al., 1984; Wray and Roeder, 1987). It may also potentate or exacerbate the pathogenicity of co-infecting pathogens and may thereby increase the severity of concomitant diseases (Potgieter, 1988; de Verdier Klingenberg, 2000). The risk for respiratory disease and increased respiratory sounds was increased on farms with poor record-keeping while the opposite was found for diarrhoea. With regard to diarrhoea, farmers who had a ‘‘less good’’ record-keeping capacity probably did not notice all the cases of diarrhoea. A similar assumption between increased milk production at the
140
G.K. Lundborg et al. / Preventive Veterinary Medicine 68 (2005) 123–143
herd level and increased disease treatment frequencies has been made by Emanuelson and Oltenacu (1998). Respiratory disease was largely diagnosed by the project veterinarians who also performed all the lung-auscultations. The farmers’ thoroughness in keeping records of calf diseases is, however, likely to be related to their thoroughness in managing the calves, which could of course have an impact on the disease risk. We did not find the sex and status of the caretaker to be associated with the herd-level morbidity. The literature reports conflicting results concerning this. Speicher and Hepp (1973) reported higher calf losses when hired staff rather than family members cared for the calves. Whereas Hartman et al. (1974), Martin et al. (1975) and Jenney et al. (1981) reported similar results, Oxender et al. (1973), James et al. (1984) and Hagstad et al. (1984) could not demonstrate statistically significant differences between animals cared for by hired staff and animals tended by the owner. James et al. (1984) found a lower mortality when the wife or children cared for the calves. Trends of better results for the owner’s mother or wife were also reported by Hartman et al. (1974) and Speicher and Hepp (1973). Waltner-Toews et al. (1986a) reported that during the summer, family members other than owner had less calf mortality than the owner him/herself and discussed how factors such as methods of grouping and a small number of farms with hired staff could affect the results. Also in the present study, only a small number of farms had hired staff. However, the fact that the capacity of keeping accurate records was found to have an effect on the disease risks may be viewed as an indicator of the caretakers’ importance. It is well known that sudden changes in the diet may result in diarrhoea. Wray and Thomlinson (1975) observed that increases in the proportion of total pathogenic Escherichia coli in the faeces of young calves often followed changes in the liquid diet. We did not find an association between diarrhoea and method of weaning. This could be due to the fact that farmers may have failed to register cases of diarrhoea at weaning, either because the period of faeces looser than normal was shorter than our stipulated 2 days or because they considered diarrhoea at weaning as normal. In the present study, no significant association was found between the cleanliness of the bedding and calf morbidity. Curtis et al. (1993) found damp bedding to be a risk factor for enteritis within 14 days after birth.
5. Conclusion The result of this study stresses the impact of environmental and management factors and gives more evidence for the importance of an infection with BVDV. The advices that could be extrapolated from this study are that calf pens should be placed separated from outer walls and kept draught-free and that keeping the herd BVDV-free is a good investment in the health of the calves.
Acknowledgements This study was financially supported by the Swedish Farmers’ Foundation for Agricultural Research, the Swedish Council for Forestry and Agricultural Research, the
G.K. Lundborg et al. / Preventive Veterinary Medicine 68 (2005) 123–143
141
Sko¨ vde and Va¨ xjo¨ Djurskyddsfo¨ reningar, Stiftelsen Skaraborgs No¨ tkreaturfo¨ rsa¨ kringsbolagsfond, Agrova¨ st, the Swedish Dairy Association and the Bro¨ derna Johnssons Forskningsfond. We would like to thank the participating farmers. We would also like to extend thanks to Kerstin Plym Forshell (formerly of Svensk Mjo¨ lk; present affiliation: Tine Norske Mejerier), Sven-Ove Olsson (of Svensk Mjo¨ lk), and Ulf Emanuelson (of Epi-Lys) for their help in designing the study, and Jan Nilsson (of the Swedish University of Agricultural Sciences) for help with the editing of data. We are also grateful to Lotta Andersson and ¨ stlund (of the Swedish University of Agricultural Sciences) and the staff at Skara Jonica O Semin for their help with collection of data, and to Gunilla Jacobsson (of the Swedish University of Agricultural Sciences) for practical data handling. Finally, we would like to thank Daniel Maizon (of the Cornell University) and Jan Hultgren (of the Swedish University of Agricultural Sciences) for valuable help with the statistical analyses.
References Alenius, S., Lindberg, A., Larsson, B., 1997. A national approach to the control of bovine viral diarrhoea virus. In: Edward, S., Paton, D.J., Wensvoort, G. (Eds.), Third ESVV Symposium on Pestivirus Infections, 19–20 September 1996, Lelystadt, The Netherlands, pp. 162–169. Andrews, A.H., 2000. Calf pneumonia costs! BVCA 8, 109–114. Barber, D.M.L., Nettleton, P.F., Herring, J.A., 1985. Disease in a dairy herd associated with the introduction and spread of bovine virus diarrhoea virus. Vet. Rec. 117, 459–464. Curtis, C.R., Erb, H.N., Scarlett, J.M., White, M.E., 1993. Path model of herd-level risk factors for calfhood morbidity and mortality in New York Holstein herds. Prev. Vet. Med. 16, 223–237. Curtis, C.R., Scarlett, J.M., Erb, H.N., White, M.E., 1988. Path model of individual-calf risk factors for calfhood morbidity and mortality in New York Holstein herds. Prev. Vet. Med. 6, 43–62. de Verdier Klingenberg, K., 2000. Enhancement of clinical signs in experimental rotavirus infected calves by combined viral infections. Vet. Rec. 147, 717–719. de Verdier Klingenberg, K., Va˚ gsholm, I., Alenius, S., 1999. Incidence of diarrhea among calves after strict closure and eradication of bovine viral diarrhea virus infection in a dairy herd. J. Am. Vet. Med. Assoc. 214, 1824– 1828. Emanuelson, U., Oltenacu, P.A., 1998. Incidences and effects of diseases on the performance of Swedish dairy herds stratified by production. J. Dairy Sci. 81, 2376–2382. Ersbøll, A.K., Rugbjerg, H., Stryhn, H., 2001. Increased mortality among calves in Danish cattle herds during bovine virus diarrhoea infection.. XIth International Conference on Production Diseases in Farm Animals, Fredriksberg, Denmark, August 12–16, pp. 77–78. Fourichon, C., Beaudeau, F., Seegers, H., 1997. Critical points related to housing and management in control programmes for calf morbidity and mortality in French dairy herds.. In: Proceedings of the Ninth International Congress on Animal Hygiene, Helsinki, Finland, August 17–21, pp. 32–35. Goldstein, H., 1995. Multilevel Statistical Models. Edward Arnold/Wiley, London/New York. Goldstein, H., Browne, W.J., Rasbach, J., 2002. Partitioning variation in multilevel models. Understanding Stat. 1, 223–241. Goldstein, H., Rasbach, J., 1996. Improved approximations for multilevel models with binary responses. J. R. Stat. Soc. 159, 505–513. Gunn, G.J., Stott, A.W., 1997. A comparison of economic losses due to calf enteritis and calf pneumonia in northern Scotland. In: Proceedings of the VIIIth International Symposium of Veterinary Epidemiology and Economics, Paris, France, July 8–11, pp. 10.06.1–10.06.3. Hagstad, H.V., Nicholson, S.S., Fulton, R.W., Springer, W.T., Cox, H.U., 1984. Influence of management practices on dairy calf mortality. Trop. Vet. 2, 123–127.
142
G.K. Lundborg et al. / Preventive Veterinary Medicine 68 (2005) 123–143
Harry, E.G., 1978. Air pollution in farm buildings and methods of control: a review. Avian Pathol. 7, 441–454. Hartman, D.A., Everett, R.W., Slack, S.T., Warner, R.G., 1974. Calf mortality. J. Dairy Sci. 57, 576–578. Houe, H., Meyling, A., 1991. Surveillance of cattle herds for bovine virus diarrhoea virus (BVDV)-infection using data on reproduction and calf mortality. Arch. Virol. Suppl. 3, 157–164. James, R.E., McGilliard, M.L., Hartman, D.A., 1984. Calf mortality of Virginia dairy herd improvement herds. J. Dairy Sci. 67, 908–911. Jenney, B.F., Gramling, G.E., Glaze, T.M., 1981. Management factors associated with calf mortality in South Carolina dairy herds. J. Dairy Sci. 64, 2284–2289. Jennings, A.R., Glover, R.E., 1952. Enzootic pneumonia in calves. II. The experimental disease. J. Comp. Pathol. 62, 6–22. Kilburn, K.H., 1967. Cilia and mucus transport as determinants of the response of the lung to air pollutants. Arch. Environ. Health 14, 77–90. Kiorpes, A., Butler, D., Dubielzieg, R., Beck, K., 1988. Enzootic pneumonia in calves: clinical and morphologic features. Comp. Cont. Educ. Pract. Vet. 10, 248–258. Kirkwood, J.K., Hubrecht, R., 2001. Animal consciousness, cognition and welfare. Anim. Welf. 10, 5–17. Larsson, B., Niskanen, R., Alenius, S., 1994. Natural infection with bovine virus diarrhoea virus in a dairy herd: a spectrum of symptoms including early reproductive failure and retained placenta. Anim. Reprod. Sci. 36, 37– 48. Martig, J., Boss, P.H., Nicolet, J., Steck, F., 1976. Etiology and predisposing factors in respiratory disease of milkfattened veal calves. Livestock Prod. Sci. 3, 285–295. Martin, S.W., Schwabe, C.W., Franti, C.E., 1975. Dairy calf mortality rate: influence of management and housing factors on calf mortality rate in Tulare County, California. Am. J. Vet. Res. 36, 1111–1114. MLwiN Help Version 3.0.02, March 2002. # Centre for Multilevel Modelling, Institute of Education. Moerman, A., Straver, P.J., de Jong, M.C., Quak, J., Baanvinger, T., van Oirshot, J.T., 1994. Clinical consequences of a bovine virus diarrhoea virus infection in a dairy herd: a longitudinal study. Vet. Q. 16, 115–119. Olsson, S.-O., Jacobsson, L., Alenius, S., Larsson, B., 1993. A voluntary control programme against infection with bovine diarrhoea virus (BVDV). Svensk Vet. Tidn. 45, 411–415. Oxender, W.D., Newman, L.E., Morrow, D.A., 1973. Factors influencing dairy calf mortality in Michigan. J. Am. Vet. Assoc. 162, 458–460. Pettersson, K., Svensson, C., Liberg, P., 2001. Housing, feeding and management of calves and replacement heifers in Swedish dairy herds. Acta Vet. Scand. 42, 465–478. Potgieter, L.N.D., 1988. Immunosuppression in cattle as a result of bovine viral diarrhea virus infection. Agric. Pract. 9, 7–14. Potgieter, L.N.D., McCracken, M.D., Hopkins, F.M., Walker, R.D., 1984. Effect of bovine viral diarrhea virus infection on the distribution of infectious bovine rhinotracheitis virus in calves. J. Am. Vet. Res. 45, 687–690. Rasbash, J., Browne, W., Goldstein, H., Yang, M., Plewis, I., Healy, M., Woodhouse, G., Draper, D., Langford, I., Lewis, T., 2002. A User’s Guide to MLwiN. Version 2.1d for Use with MLwiN 1.10. Centre for Multilevel Modelling, Institute of Education, London. Roth, J.A., Kaeberle, J.W., 1983. Suppression of neutrophil and lymphocyte function induced by a vaccinal strain of bovine viral diarrhea virus with and without the administration of ACTH. Am. J. Vet. Res. 44, 2366–2372. Simensen, E., Norheim, K., 1983. An epidemiological study of calf health and performance in Norwegian dairy herds. IV. Factors affecting morbidity and performance. Acta Agric. Scand. 33, 65–74. Speicher, J.A., Hepp, R.E., 1973. Factors associated with calf mortality in Michigan dairy herds. J. Am. Vet. Med. Assoc. 162, 463–466. Svensson, C., Lundborg, K., Emanuelson, U., Olsson, S.-O., 2003. Morbidity in Swedish dairy calves from birth to 90 days of age and individual calf-level risk factors for infectious diseases. Prev. Vet. Med. 58, 179–197. Tra˚ ve´ n, M., Alenius, S., Fossum, C., Larsson, B., 1991. Primary bovine viral diarrhoea virus infection in calves following direct contact with a persistently viraemic calf. J. Vet. Med. B 38, 453–462. Virtala, A.M.K., Gro¨ hn, Y.T., Mechor, G.D., Erb, H.N., 1999. The effect of maternally derived immunoglobulin G on the risk of respiratory disease in heifers during the first 3 months of life. Prev. Vet. Med. 39, 25–37. Waltner-Toews, D., Martin, S.W., Meek, A.H., 1986a. Dairy calf management, morbidity and mortality in Ontario Holstein herds. III. Association of management with morbidity. Prev. Vet. Med. 4, 137–158. Waltner-Toews, D., Martin, S.W., Meek, A.H., McMillian, I., 1986b. Dairy calf management, morbidity and mortality in Ontario Holstein herds. I. The data. Prev. Vet. Med. 4, 103–124.
G.K. Lundborg et al. / Preventive Veterinary Medicine 68 (2005) 123–143
143
Wathes, C.M., 1998. Environmental control in pig housing. In: Done, S., Thomson, J., Varley, M. (Eds.), XVth International Pig Veterinary Society Congress, Birmingham. Nottingham University Press, Nottingham, Great Britain, pp. 257–265. Wathes, C.M., Jones, C.D.R., Webster, A.J.F., 1983. Ventilation, air hygiene and animal health. Vet. Rec. 113, 554–559. Wray, C., Roeder, P.L., 1987. Effect of bovine virus diarrhoea–mucosal disease virus infection on salmonella infection in calves. Res. Vet. Sci. 42, 213–218. Wray, C., Thomlinson, J.R., 1975. Factors influencing the occurrence of colibacillosis in calves. Vet. Rec. 96, 52–56.
Herd-level risk factors for infectious diseases in Swedish dairy calves aged 0–90 days G.K. Lundborga, E.C. Svenssona,*, P.A. Oltenacub a
Department of Animal Environment and Health, Faculty of Veterinary Medicine and Animal Science, Swedish University of Agricultural Sciences, P.O. Box 234, SE-532 23 Skara, Sweden b Department of Animal Science, Cornell University, Ithaca, NY 14853, USA
Received 19 February 2004; received in revised form 18 November 2004; accepted 30 November 2004
Abstract The effect of environmental factors and management routines on the risk of diarrhoea, respiratory disease and other infectious diseases was investigated in 3081 heifer calves 0–90 days old in 122 Swedish dairy herds. The farmers kept records on cases of diseases in their heifer calves and in addition, project veterinarians clinically examined all calves every 2–3 months. At each visit, the veterinarians also measured the ammonia concentration and relative air humidity in the housing facilities for the calves. The cleanliness of the animals and their environment was recorded as a measure of the hygienic status of the farm. The presence or absence of draught (i.e. wind velocity > 0.5 m/s) was recorded twice during the study period. The effect of these factors, as well as the placing of the calf pens, the nature of the pen walls, air volume per animal, management factors (such as the status of the caretaker and feeding routines) and presence or absence of a bovine viral diarrhoea virus (BVDV) infection in the herd, was evaluated by means of a two-level variance component logistic model. The placing of calf pens along an outer wall was significantly associated with the risk of diarrhoea (odds ratio (OR): 1.92, P < 0.01). The risk for respiratory disease was significantly associated with an ammonia concentration below 6 ppm (OR: 0.42, P < 0.05) while the odds ratio for moderately to severely increased respiratory sounds was significantly associated with a BVDV infection in the herd (OR: 2.39, P < 0.05) and draught (OR: 3.7, P < 0.02). Absence of
* Corresponding author. Tel.: +46 511 67205; fax: +46 511 67204. E-mail address: [email protected] (E.C. Svensson). 0167-5877/$ – see front matter # 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.prevetmed.2004.11.014
124
G.K. Lundborg et al. / Preventive Veterinary Medicine 68 (2005) 123–143
draught was significantly associated with the risk for infectious diseases other than diarrhoea and respiratory disease (OR: 0.42, P < 0.01). # 2004 Elsevier B.V. All rights reserved. Keywords: Ambient environment; Bovine viral diarrhoea virus (BVDV); Dairy calf; Diarrhoea; Herd-level morbidity; Management; Respiratory disease; Risk factors
1. Introduction In the countries within the European Union, the interest of the general public in animal welfare has increased during the past decades (Kirkwood and Hubrecht, 2001). For the consumer, the conditions the animals are kept in are an increasingly important aspect of the quality of dairy products, making this a valuable sales argument for the dairy industry. To many consumers, the welfare of the dairy calf is as important as is that of the adult animal. However, for the farmer, the calf often comes second to the lactating cow. Suboptimal management of and a suboptimal environment for dairy calves are not only an animal welfare issue, but also have a direct economic impact on the farming enterprise as they are likely to result in a higher calf morbidity and lower growth rates. Gunn and Stott (1997) estimated the average loss per calf at risk in a herd with an outbreak of enteritis, accounting for calf mortality, loss in calf value, expenses for extra work, possible veterinary treatment and extra feed to compensate for the retained growth, to be approximately £33. They found the corresponding loss per calf at risk in a herd with an outbreak of pneumonia to be approximately £21 (Gunn and Stott, 1997). Andrews (2000) calculated the average extra rearing cost to be £43 for a diseased calf in a herd with an outbreak of pneumonia. Although management and the environment are known to play an important role in the clinical outcome of the multi-factorial infectious diseases that are the main health disorders in calves, knowledge about the effects of individual factors is still limited. Pneumonia has been associated with poor climatic conditions (Kilburn, 1967; Harry, 1978; Kiorpes et al., 1988). However, few studies have objectively measured air quality parameters such as ammonia concentration and relative humidity and examined their effect on the risk of respiratory disease in calves. Similarly, draught is often mentioned as a risk factor for calf morbidity, but we have not come across any study that more objectively examined its effect. The literature offers conflicting results about the effects on calf health of housing calves separately (Simensen and Norheim, 1983; Fourichon et al., 1997; Virtala et al., 1999) and of the sex and status of the calf caretaker (Oxender et al., 1973; Speicher and Hepp, 1973; Hartman et al., 1974; Jenney et al., 1981; Hagstad et al., 1984; James et al., 1984; WaltnerToews et al., 1986a), and no firm conclusions have been drawn about the relationship between calf health and herd size (Oxender et al., 1973; Speicher and Hepp, 1973; Hartman et al., 1974; James et al., 1984; Waltner-Toews et al., 1986a; Curtis et al., 1988). With regard to the effects of management and feeding procedures, the literature is surprisingly sparse. Introduction of bovine viral diarrhoea virus (BVDV) into the dairy herd has been associated with impaired calf health (Barber et al., 1985; Houe and Meyling, 1991; Larsson
G.K. Lundborg et al. / Preventive Veterinary Medicine 68 (2005) 123–143
125
et al., 1994; Moerman et al., 1994) and de Verdier Klingenberg et al. (1999) demonstrated a decreased incidence of calf diarrhoea after strict closure and eradication of BVDV infection in a dairy herd. However, the reports mentioned above all cover the situation in a single or a few herds only and there are few large-scale studies on the effect of BVDV status on calf morbidity. A large-scale study performed by Ersbøll et al. (2001) reported an increased calf mortality rate in herds with a BVDV infection. The aim of the present paper was to describe the herd-level morbidity and evaluate the effect of some potential herd-level risk factors on the incidence of infectious diseases in 0– 90-day-old dairy calves reared under Swedish conditions.
2. Material and methods 2.1. Animals The study was performed in all heifer calves born on 122 farms in the county of Skaraborg in south-western Sweden during 1998, in total 3081 calves. All the farms were enrolled in the official milk recording programme. All farmers had a herd size of between 28 and 94 cows (median 47.5 cows), housed their young calves in individual or group pens and were considered by the large animal practitioners and farm advisors in the region to be capable of keeping thorough records of their calves’ performance. A more detailed description of the selection of the participating farms can be found in a previous paper (Svensson et al., 2003). After a statistical comparison between the 122 farms and a random sample of 877 Swedish dairy herds of similar size (Pettersson et al., 2001; Svensson et al., 2003), the farms were considered a reasonably representative sample of Swedish dairy farms with 28–94 cows. They were mostly run exclusively by family members, but 28% had one or two employees. The population of heifer calves monitored was dynamic throughout the study period. Calves entered the monitoring phase of the study at birth and left at 90 days of age. Altogether, 108 (3.5%) calves died or were slaughtered and 26 (0.8%) were sold or lost to the study before 90 days of age. Calves were generally kept indoors and housed in single pens, small group pens for between 3 and 8 calves, with manual feeding of the milk, or in large group pens for 6–30 calves with an automatic milk-feeding system. Sixty-three percent of the calves were removed from their mothers immediately after birth; the rest were allowed to stay with their mothers for 8 h–4 days. Almost all calves received colostrum from their own mother (98%). Pooling of colostrum is very rarely used in Sweden. Colostrum exclusively from the first milking was used for the calves’ first two meals by 36% of the farmers. Calves received whole milk until weaning in 45% of the herds and milk replacements in 45% of the herds. In the remaining herds, they were fed a combination of milk replacement and whole milk. The milk volume given ranged from 3 to 8 L (median 5.0 L) per day. The calves were weaned at a median of 9.0 weeks of age (20–80th percentiles: 8–11 weeks). They were given free access to hay. Concentrates, mainly as pelleted calf feed or as crushed grain solely or in combination with protein feed, were given starting from day 1 to day 112 (median day 7). The concentrates were given ad libitum in 88% of the herds.
126
G.K. Lundborg et al. / Preventive Veterinary Medicine 68 (2005) 123–143
2.2. Data collection The farmers were requested to keep records on cases of disease (treated as well as untreated) in their heifer calves on special individual health cards. On the health cards, they also continuously recorded for each calf date of birth, place of birth (BIRTHPLACE; individual maternity pen, tie stall, cubicle or group maternity pen, or pasture), time of birth (BIRTHTIME; day or night), breed (BREED), method of feeding colostrum (COLFEED; suckling or by bucket/nipple feeder), origin of colostrum (COLORIGIN; a primiparous cow, a second calving cow, a cow that has calved three or more times), hours from birth to first colostrum feeding (COLTIME), housing system used for the calves (HOUSING; individual pen, small-group pen with manual feeding of the milk or large-group pen with an automatic milk feeding system) and supervision of calving (SUPERVIS; yes or no). For a more detailed description and for the distributions of these calf-level variables, see Svensson et al. (2003). Based on our definitions of respiratory disease and diarrhoea, the farmers recorded these two diseases directly on the individual health cards. ‘‘Diarrhoea’’ was defined as a faecal consistency that for 2 days or more was softer than is normally observed in cattle of the age in question. ‘‘Respiratory disease’’ was defined as coughing or sneezing for more than 2 days, as severely increased respiratory sounds at lung auscultation or as moderately increased respiratory sounds together with additional signs, such as coughing or nasal discharge. For all other disease conditions, the farmers recorded the symptoms and the project veterinarians on their visits (every second month) determined the diagnosis, based on mutual definitions. On these visits, the project veterinarians also clinically examined all calves, auscultated their lungs, checked the notes kept by the farmers and recorded prevalent diseases not detected by the farmers in the health records. Recorded infectious diseases were divided by the veterinarians into the following categories: arthritis, omphalophlebitis/umbilical abscess, weak calf syndrome and ‘‘other infectious disease’’ (e.g. abscesses and ringworm). ‘‘Arthritis’’ was defined as swelling of one or more joints accompanied by lameness and fever. ‘‘Omphalophlebitis/umbilical abscess’’ was defined as a warm swelling or abscess formation associated with the navel. Weak calf syndrome was diagnosed if an animal had inappetence and dullness during at least 2 days, with or without an increased body temperature, but lacked other symptoms of disease. Lungauscultation findings were ranked as normal respiratory sounds or as mildly, moderately or severely increased respiratory sounds. At each farm visit, the veterinarians also measured the ammonia concentration (AMMONIA) and the relative air humidity (HUMIDITY) in the housing facilities for the calves. The same measure points were used at all visits: at the nose level of standing animals in the most central pen used for calves 0–90 days of age. If there were several types of calf pens in the herd (e.g. calves housed in group pens were most often kept in single pens during their first weeks of life and singled-housed calves were sometimes moved to group pens before 90 days of age), one measure point for each pen type was used. Recordings were made on visits from January 1998 to April 2000, so that in total 8–12 measurements were made for each pen type and herd, reflecting conditions during different months of the year. The ammonia concentration was measured using a mobile Dra¨ ger1 gas monitor Pac III (Dra¨ gerwerk AG, Lu¨ beck, Germany), while relative humidity was
G.K. Lundborg et al. / Preventive Veterinary Medicine 68 (2005) 123–143
127
measured using a Testo1 term hand instrument (Testoterm GmbH & Co., Lenzkirch/ Schwarzwald, Germany). The cleanliness of the animals (ANIMCLEAN) and of the bedding (BEDCLEAN), the concentrate and water buckets (FBUCKCLEAN and WBUCKCLEAN) and the pen walls (PENCLEAN) were recorded at each visit as measures of the hygienic status of the farm. On each of the first five visits to a farm, the two next youngest animals were chosen until 10 individuals per farm had been selected. During each farm visit, the project veterinarian monitored the cleanliness of each of these 10 animals as well as of the walls and bedding of their present pen. Their concentrate and water buckets were also examined and their cleanliness scored. The animals were given one cleanliness score for each of eight different body parts (front claws and hind claws, front legs below the carpus, hind legs below the hock, brisket, thighs, belly and tail). The scores ranged from 0 to 100% and represented the percentage of the body part that was contaminated with faeces. A value of 0% described a clean area, 5–30% a mildly contaminated, >30–70% a moderately contaminated and >70% a heavily contaminated body area. The cleanliness of pen walls, bedding, and the concentrate and water buckets were scored similarly from 0 to 100%. Before the start of the study and twice during the study, simultaneous monitoring was made by the three project veterinarians in order to standardise their scales. Once in the summertime and once in the wintertime, presence or absence of draught (DRAUGHT) was determined in each of the calf pens on the farms using a smoke bottle (Titranetetrachloride RFA; AB Regin, Sweden). Draught was considered present (score 1) if wind velocity exceeding 0.5 m/s was detected in any of the pens. The project veterinarians also measured the air volume in the calf premises and calculated the number and category of animals per building and the number of animals per pen. They also recorded information on placing of the calf pens within the building and nature of pen walls (e.g. possibilities for direct nose contact between calves from different pens). Based on their experiences during the study, the three project veterinarians jointly graded the capability of the farmers (REGCAPACITY) to keep accurate records of diseases in their calves as ‘‘less good’’, ‘‘good’’ or ‘‘excellent’’. Data on feeding and management routines were collected by the project veterinarians by means of interviews with the farmers at farm visits. The farmers were asked about the routines for cleaning the milk buckets (BUCKRINSE), the storage of the milk buckets between meals (BUCKSTORE), the use of the milk buckets for the calves (BUCKUSE), the routines for cleaning the concentrate buckets (CONBUCKCLEAN), whether concentrate not eaten was removed (CONCREMOVED), the routines for dosing the milk powder (MILKDOSE), milk type used (MILKTYPE), whether they gave the cows ADE-vitamins before calving (VITAMIN), whether they heated the whole milk by adding hot water (WATERMILK) and weaning routines (WEANHOW). For a description of the categories within each variable, see Table 1a–c. The farmers were also interviewed about feeding routines and routines at calving and weaning, as well as about the status of the caretakers and the number of animals bought annually (see Table 1a–c). Data on the BVDV status of the herds were collected from the Swedish BVDV control programme. According to the control programme, a herd is certified as BVDV-free if two samples (i.e. samples of bulk milk, pooled milk samples from primiparous cows or individual blood samples) collected with a 7 months’ interval are seronegative for the virus.
128 Table 1 (a) Management variables, (b) feeding hygiene-related variables and (c) feeding-related variables constructed from information gathered through interviews with farmers, and (d) variables constructed from observations or measurements made by project veterinarians during visits on 122 dairy farms in south-western Sweden
(a) Management variables BUYANIM Number of animals bought annually
CALVINGPEN Percentage of times when the calving pen was cleaned between consecutive calvings CARETAKER Number of persons involved in the care of the calves
CHORE Percentage of cases when the feeding of the calves was done at the same time as another chore COWCALF Proportion of the calves that were held with the dam sometime after birth COWNO Number of cows in the herd
Categories
Number of herds
Included in multivariate analysesa
1 = 5 2 = 1 to <5 3=0
11 13 98
I
1 = >75 2 = >25 to <75 3 = 25
39 12 8
R, A, I
1 = 3 2=2 3=1
21 60 41
D, I
1 = >75 2 = >25 to <75 3 = 25
20 16 86
R, A
1=0 2 = >0 to <1 3=1
60 44 18
D, R, A, I
1 = 59 2 = 40 to <59 3 = <40
31 64 27
D, R, A, I
G.K. Lundborg et al. / Preventive Veterinary Medicine 68 (2005) 123–143
Variable
EMPLOYEE Percentage of calf caretaking performed by an employee
VITAMIN Whether cows were routinely supplied with ADE vitamins before calving or not WEANAGE Age at weaning (weeks)
WEANHOW Routines for weaning
(b) Feeding hygiene-related variables BUCKRINSE Routines for rinsing milk buckets
BUCKSTORE Routines for storing milk buckets between the meals
11 17 94
1 = >75 2 = >25 to <75 3 = 25
32 30 60
D, R, A, I
1 = yes 2 = no
23 99
D, R, A, I
1 = 10 2 = >8 to <10 3 = 8
39 29 54
1 = weaning abruptly 2 = weaning by reducing milk meals to one per day 3 = weaning by offering less milk per meal 4 = weaning by adding increasing amounts of water to the milk
28 27
1 = buckets rinsed after every meal 2 = buckets rinsed less often than after every meal
100 20
43 24
22 48 49
D 129
1 = combination 2 = stored separately 3 = stacked when stored
D, R, A
G.K. Lundborg et al. / Preventive Veterinary Medicine 68 (2005) 123–143
FEMALE Percentage of calf caretaking performed by a woman
1 = >75 2 = >25 to <75 3 = 25
130
Table 1 (continued ) Variable
CONBUCKCLEAN Routines for cleaning concentrate buckets
CONCREMOVED Whether concentrates not eaten are removed or not (c) Feeding-related variables COLFIRSTMILK Percentage of cases in which colostrum from the first milking was used for more feeds than the first COLLITRES Litres of colostrum fed on the first day
Number of herds
1 = one unique bucket per calf 2 = one bucket per calf and meal 3 = one bucket for several calves during the same meal
15 30 72
1 = buckets cleaned every second week 2 = buckets cleaned once a month 3 = buckets cleaned when calf leaves the pen 4 = buckets cleaned less often than every 7th week
17
Included in multivariate analysesa
29 27 31
1 = yes 2 = no
92 17
1 = >75 2 = >25 to <75 3 = 25
53 17 52
D, R, A, I
1 = 8 2 = >4 to <8 3 = 4
11 75 36
D, R, A, I
G.K. Lundborg et al. / Preventive Veterinary Medicine 68 (2005) 123–143
BUCKUSE Routines for use of milk buckets
Categories
CONCAGE Age at which the calves were first offered concentrates (weeks)
MILKDOSE Dosage of milk powder in milk replacement
MILKTYPE Type of milk
TEMPCHECK Percentage of meals at which the temperature of the milk was checked WATERMILK Whether the whole milk was heated by adding hot water or not WATERAGE Age at which the calves were first offered water (weeks)
15 34 73
D
1 = >7 2 = >1 to 7 3 = 1
11 49 60
D
1 = according to manufacturer’s instructions 2 = different from manufacturer’s instructions
43 15
1 = combination 2 = whole milk 3 = milk replacement
9 56 56
1 = >75 2 = >25 to <75 3 = 25
75 12 28
D
1 = yes 2 = no
14 40
R, A
1 = >4 2 = 2 to <4 3 = <2
47 28 47
G.K. Lundborg et al. / Preventive Veterinary Medicine 68 (2005) 123–143
HAYAGE Age at which the calves were first offered hay (days)
1 = >2 2 = >1 to 2 3 = 1
(d) Variables constructed from observations or measurements made by project veterinarians 131
132
Table 1 (continued ) Variable
ANIMCLEAN Cleanliness of the animals (percentage of the animal surface contaminated with faeces) AMMONIA Ammonia concentration (ppm) BEDCLEAN Cleanliness of the bedding (percentage of the bedding contaminated with faeces) BUILDING Housing of calves in relation to older cattle
BVDV BVDV-statusb
CONTACT Possibilities for nose contact with calves from another pen
1 = >20 2 = 20
Number of herds 60
Included in multivariate analysesa R, A
61 1 = >30 2 = 30
8 114
1 = 6 2 = <6
38 64
1 = >30 2 = 30
28 94
1 = calves housed in the same building section as cows and/or older young stock 2 = calves housed separately 1 = BVDV introduced into the herd during the study period 2 = BVDV present in the herd at the start of the study period 3 = BVDV not present in the herd 1 = no 2 = yes
R, A, I
103
19 12 32 78 104 17
D, R, A, I
G.K. Lundborg et al. / Preventive Veterinary Medicine 68 (2005) 123–143
AIRVOLUME Air volume (m3 air/animal unit) (calves = 0.2 units; young stock >6 months = 0.5; cow or heifer near calving = 1.0 units)
Categories
DRAUGHT Draught (defined as a wind velocity exceeding 0.5 m/s detected in any pen)
HUMIDITY Relative air humidity (%) PENCLEAN Cleanliness of the pen (%) PENPLACE Placing of the calf pens
REGCAPACITY Farmers’ capability of keeping accurate records
WBUCKCLEAN Cleanliness of the water bucket (percentage of buckets that were considered clean) a b
17
D, R, A, I
53
1 = >30 2 = 30
47 64
1 = >80 2 = 80
13 109
1 = >30 2 = 30
55 67
D, R, A, I
1 = against an outer wall 2 = separated from the outer walls 3 = combinations
29 42 50
D, R, A, I
1 = less good 2 = good 3 = excellent
20 82 20
D, R, A, I
1 = >30 2 = 30
8 43
D, I
52
D, I
R, A, I
Variables examined for their effects on risk of diarrhoea (D), respiratory disease (R), increased respiratory sounds (A) and other infectious disease (I). Information retrieved from the Swedish official BVDV (bovine viral diarrhoea virus) control programme.
G.K. Lundborg et al. / Preventive Veterinary Medicine 68 (2005) 123–143
FBUCKCLEAN Cleanliness of the concentrate buckets (percentage of the buckets that were considered clean)
1 = 1 (draught recorded both in summer and winter) 2 = >0 to <1 (draught recorded in summer or winter) 3 = 0 (no draught recorded)
133
134
G.K. Lundborg et al. / Preventive Veterinary Medicine 68 (2005) 123–143
For a closer description of the eradication programme, see Olsson et al. (1993) and Alenius et al. (1997). 2.3. Editing of data and statistical analysis In cases where results from more than one lung-auscultation were recorded in a calf prior to 91 days of age, the last recording was used. No findings recorded in calves younger than 31 days were used in the study. The findings at lung auscultation were dichotomised into normal to mildly increased respiratory sounds (0) and moderately to severely increased respiratory sounds (1). A variable SEASON was constructed based on the date of birth. Birth dates during the period from 1 May to 31 August was defined as summer dates, 1 September to 30 November defined as autumn dates and 1 December to 30 April as winter dates. Herd AMMONIA and HUMIDITY values were constructed from mean values of all recordings in the herd. If more than one pen type was used for calves in the herd, a mean value was first calculated for each pen type. An overall weighted mean was thereafter calculated based on the percentage of weeks from day 0 to day 90 that the calves usually spent in each pen type. Herd DRAUGHT scores were calculated as the mean of the summer and winter scores in each herd (scores between 0 and 1). Herd-level scores were also constructed for the five cleanliness parameters. For each of the 10 individuals monitored per farm, median body part scores were first constructed based on all scores recorded for the particular body part. Thereafter, the median body part scores were each used to construct median animal scores, calculated as the median score of the three body parts with the highest scores (if hind and front claws were among these three body parts, only the highest claw score was included). An ANIMCLEAN score for the herd was calculated as the median of all animal scores in the herd. PENCLEAN, BEDCLEAN, FBUCKCLEAN and WBUCKCLEAN scores for each of the herds were calculated similarly. Continuous predictor variables were transformed to categorical variables; cut-off points were based on biological knowledge, practical common routines or quartiles. The categories and their distributions can be seen in Table 1a–d. The herd-level variables were initially screened in univariate analyses for each of the four outcome variables: diarrhoea, respiratory disease, ‘‘other infectious disease’’ and moderately to severely increased respiratory sounds, using the procedure PROC LOGISTIC in SAS (SAS Institute Inc., 2000). All herd-level variables that were associated with an outcome variable at P 0.2 using the F-test were then transferred to the multi-level statistical software MLwiN (MLwiN, 2002; Rasbash et al., 2002). In addition, we considered calf-level factors found to significantly affect the incidence risk of the outcome variables in a previous study based on the same material (Svensson et al., 2003): BIRTHPLACE, BREED, BIRTHTIME, COLFEED, COLORIGIN, COLTIME, HOUSING, SEASON and SUPERVIS. As with herd-level variables, calf-level variables associated with the outcome at P 0.2 in a univariate analysis using the F-test were transferred to MLwiN. For each binary outcome, a two-level (calf; herd) variance components logistic model was built, using second-order penalised quasi-likelihood estimations (Goldstein and
G.K. Lundborg et al. / Preventive Veterinary Medicine 68 (2005) 123–143
135
Rasbach, 1996) with the restricted iterative generalised least-squares method (Goldstein, 1995). The models were defined by the following equations: Yij ¼ pij þ eij logitðpij Þ ¼ b0j þ
X
bm Xmij
b0j ¼ b0 þ uj where Yij is the disease status of calf i in herd j, pij the probability that calf i in herd j was diseased, eij a random term at the calf-level, b0j the intercept of herd j, bm the regression coefficients expressing effects of the included predictors Xmij, b0 the mean intercept among all herds, and uj is the random term at the herd-level. The final models were built by a manual forward stepwise procedure, including one factor at a time and excluding factors that were not significant, according to the approximate two-sided Wald (large-sample joint chi-squared) test (P 0.05) before proceeding. Finally, biologically relevant first-order interactions were tested one at a time. No significant (Wald P 0.05) interactions were found. The selection of variables to be included was partly based on subjective judgement, favouring model simplicity and taking confounding factors into account; thus, in spite of the increased risk for spurious associations, a few cases of borderline significant effects (0.05 < P 0.10) were included if they were found to act as confounders and stabilise the model upon inclusion. A binomial distribution was assumed, constraining the residual variance to 1, but the degree of overdispersion was checked by re-running the models without the constraint. Level 2 residuals were checked graphically. Odds ratios (ORs) were calculated from the final parameter estimates. The proportion of unexplained variation residing at the herd-level was calculated as herd variance/(herd variance + p2/3), according to the latent-variable method (Goldstein et al., 2002). The total numbers of observations at herd as well as at calf level are given in Table 2a–d.
3. Results 3.1. Herd-level morbidity The median total herd disease incidence risk was 21.6%. However, the incidence risk differed considerably between the herds and ranged from 0.0 to 57.6%, the central range (20–80th percentiles) being 10.2–34.1%. In two herds, no morbidity was recorded; in 22 herds, no diarrhoea was recorded and in 55 herds, no cases of respiratory disease were recorded. The corresponding numbers for arthritis, other infectious disease and omphalophlebitis were 106, 87 and 93 herds, respectively. The herd disease incidence risks for the most common diseases are presented in Table 3. 3.2. Risk factors The risk for calves of suffering from diarrhoea was significantly (P < 0.01) associated with the placing of the calf pens against an outer wall, compared to calves in herds in which
136
G.K. Lundborg et al. / Preventive Veterinary Medicine 68 (2005) 123–143
Table 2 Effects of herd-level factors on incidence of (a) diarrhoea, (b) respiratory disease, (c) moderately to severely increased respiratory sounds, (d) infectious diseases other than diarrhoea and respiratory disease, from birth to 90 days of age in calves in 115–121 dairy herds in south-western Sweden from January 1998 to March 1999 Variablea
Levels
(a) Diarrhoea Intercept b,c
b
S.E. (b)
OR
95% CI (OR)
P
2.3
0.42
–
–
–
BREED
SRB SLB SLB SRB Cross-breed beef
0 0.47 0.25 0.88
– 0.17 0.48 0.78
– 0.63 0.78 0.41
– 0.45–0.87 0.30–2.00 0.09–1.91
<0.05
COLORIGINb
First-lactation cow Second-lactation cow Third or more lactation cow
0 0.53 0.31
– 0.18 0.15
– 0.59 0.73
– 0.41–0.84 0.55–0.98
<0.01
PENPLACE
Against the outer wall Separated from the outer wall Combinations
0 0.65 0.62
– 0.24 0.23
– 0.52 0.54
– 0.33–0.84 0.34–0.84
<0.01
REGCAPACITY
Less good Good Excellent
0 0.26 0.64
– 0.26 0.31
– 1.30 1.90
– 0.78–2.16 1.03–3.48
<0.10
TEMPCHECK
>25 to <75% of meals >75% of meals 25% of meals
0 0.39 0.70
– 0.37 0.33
– 1.48 2.01
– 0.72–3.05 1.05–3.85
<0.10
0.51
–
–
–
0 0.87
– 0.37
– 2.39
– 1.16–4.93
<0.05
(b) Respiratory disease Intercept
3.0
AMMONIA
6 ppm <6 ppm
HOUSINGd
Large-group pen Single pen Small-group pen
0 0.84 0.87
– 0.33 0.41
– 0.43 0.42
– 0.23–0.82 0.19–0.93
<0.05
REGCAPACITY
Less good Good Excellent
0 0.93 0.89
– 0.39 0.52
– 0.39 0.41
– 0.18–0.85 0.15–1.14
<0.10
SEASONd
Summer Autumn Winter
0 0.90 0.66
– 0.23 0.22
– 2.46 1.94
– 1.57–3.86 1.26–2.98
<0.001
(c) Moderately to severely increased respiratory sounds Intercept
0.15
0.80
–
–
–
0 0.87 0.03
– 0.36 0.59
– 2.39 1.03
– 1.16–4.93 0.32–3.26
<0.05
Detected in summer and winter Detected in summer or winter Not detected
0 0.41 1.30
– 0.46 0.50
– 0.66 0.27
– 0.27–1.63 0.10–0.73
<0.02
Large-group pen Single pen Small-group pen
0 1.70 2.00
– 0.38 0.51
– 0.18 0.14
– 0.08–0.37 0.05–0.37
<0.001
BVDV
Not present in the herd Present in the herd at start Introduced during study
DRAUGHT
HOUSINGe
G.K. Lundborg et al. / Preventive Veterinary Medicine 68 (2005) 123–143
137
Table 2 (Continued ) Variablea
Levels
b
S.E. (b)
OR
95% CI (OR)
P
REGCAPACITY
Less good Good Excellent
0 0.89 1.50
– 0.40 0.62
– 0.41 0.22
– 0.19–0.90 0.07–0.75
<0.05
WEANHOW
Weaning Weaning Weaning Weaning
0 0.69 1.10 1.10
– 0.48 0.55 0.49
– 0.50 0.33 0.33
– 0.20–1.29 0.11–0.98 0.13–0.87
<0.10
–
–
–
by adding more water squarely by reducing no of meals by reducing volume/meal
(d) Infectious diseases other than diarrhoea and respiratory disease Intercept 4.0 0.44 f
BIRTHTIME
Day Night
0 0.42
– 0.21
– 1.52
– 1.01–2.30
<0.05
DRAUGHT
Detected in summer and winter Detected in summer or winter Not detected
0 0.15 0.87
– 0.40 0.39
– 1.16 2.39
– 0.53–2.54 1.11–5.13
<0.01
COLORIGINf
Third or more lactation cow Second-lactation cow First-lactation cow
0 0.51 0.08
– 0.25 0.25
– 1.67 1.08
– 1.02–2.72 0.66–1.76
<0.10
PENPLACE
Against the outer wall Separated from the outer wall Combinations
0 0.69 0.09
– 0.31 0.29
– 0.50 0.92
– 0.27–0.92 0.52–1.62
<0.10
a
Abbreviations found in Table 1a–d. Extra explanatory calf-level variable, n = 2849. c SRB: Swedish Red and White breed; SLB: Swedish Holstein; SLB SRB: all cross-breeds between Swedish Red and Whites and Swedish Holsteins; Cross-breed beef: cross-breeds with a dam of Swedish Red and Whites or Swedish Holsteins and a sire of beef breed. d Extra explanatory calf-level variable, n = 3037. e Extra explanatory calf-level variable, n = 2049. f Extra explanatory calf-level variable, n = 2923. b
calf pens were separated from an outer wall (OR: 1.92) (Table 2a). A decreased risk for calves of suffering from respiratory disease was found to be significantly (P < 0.05) associated with a higher ammonia level (OR: 0.42) (Table 2b). The OR for moderately to severely increased respiratory sounds was significantly associated with BVDV, DRAUGHT and REGCAPACITY (Table 2c). A BVDV infection present in the herd at the start of the study increased the risk for the calves of suffering from moderately to severely increased respiratory sounds (OR: 2.39, P < 0.05) compared to calves in BVDVfree herds. The presence of draught also increased the risk for the calves of suffering from moderately to severely increased respiratory sounds (OR: 3.7, P < 0.02) compared to calves in herds in which draught was not detected. Calves in herds in which the farmer had an excellent record-keeping capacity had a decreased risk of suffering from moderately to severely increased respiratory sounds (OR: 0.22, P < 0.05). The absence of draught was significantly associated with an increased risk for the calves of suffering from an infectious disease other than diarrhoea and respiratory disease (OR: 2.39, P < 0.01) (Table 2d).
138
G.K. Lundborg et al. / Preventive Veterinary Medicine 68 (2005) 123–143
Table 3 Herd incidence risks (% calves/herd) of the most common diseases in 0–90-day-old calves from 122 dairy herds in south-western Sweden from January 1998 to March 1999
Arthritis Diarrhoea Increased respiratory sounds Omphalophlebitis Respiratory disease Other infectious diseases
Median (%)
Range (%)
20–80th percentile (%)
0.0 7.8 14.3 0.0 3.0 0.0
0.0–5.6 0.0–39.4 0.0–72.4 0.0–17.6 0.0–51.7 0.0–36.4
0.0–0.0 2.3–17.6 5.6–29.7 0.0–3.0 0.0–11.4 0.0–4.3
There was a tendency (P < 0.10) for the risk of diarrhoea to be increased for calves in herds in which the temperature of the milk was checked at less than 20% of the feeding times compared with calves in herds where the temperature check was performed 20–70% of the occasions. There was a tendency (P < 0.10) that the risk of finding moderately to severely increased respiratory sounds was greater in herds in which the weaning practice was to add increasing amounts of water to the milk compared with herds that weaned their calves by serving less and less milk per meal. We did not find an association between diarrhoea and method of weaning. In the present study, no significant association was found between the cleanliness of the bedding and calf morbidity.
4. Discussion 4.1. Herd-level morbidity The total herd-level morbidity found in the present study was of a similar level as that previously reported by Waltner-Toews et al. (1986b) and, similarly to that study, most herds had a fairly low disease incidence risk while a few herds had a high incidence risk. The skewed distribution may partly be due to the contagious nature of infectious diseases, but may also reflect differences in management and the skills of the herdsman. 4.2. Risk factors In several studies, high ammonia levels have been suggested to increase the risk of respiratory disease through irritation of the respiratory epithelium and decreasing the mucociliary flow (Kilburn, 1967; Harry, 1978; Kiorpes et al., 1988). Our data suggest a reversed relationship, which is difficult to explain. The ammonia levels recorded seem reasonable in relation to other ambient environment factors measured. However, the highest ammonia concentrations were found in pens with slatted floors, to which calves were moved only shortly before 90 days of age. For ammonia to have an effect on the level of respiratory disease, a longer exposure time may be needed, extending beyond the study period. The fact that the herd means included measurements in pens used only during the
G.K. Lundborg et al. / Preventive Veterinary Medicine 68 (2005) 123–143
139
last part of the study period may therefore have obscured the results. Our examination of the data shows that this cannot explain the unexpected result. However, the observed effect of ammonia may have been confounded by some unknown and unrecorded factor. The effects of season and caretaker were included in the model. AMMONIA showed a fairly small variation (range: 0.3–13.5 ppm; 20–80th percentiles: 3.3–6.7 ppm), which may have affected the results. Martig et al. (1976) recommended a maximum of 25 ppm ammonia for calves while Wathes (1998) recommended a 20 ppm limit for pigs. Both values far exceed the allowed level, according to the Swedish animal welfare legislation, which is 10 ppm. Another finding which is equally difficult to explain is the significant association between the detection of draught and the decreased risk of other infectious disease in a herd. Calves in herds where draught was detected were found to suffer an increased (OR: 3.7) risk of having moderately to severely increased respiratory sounds at lung auscultation, compared with calves in herds without draught. Wathes et al. (1983) defined draught as an air speed faster than 0.3 m/s, which leads to enhanced heat losses and chilling in cold weather. Martig et al. (1976) suggested that for the well being of the animals, the air speed should not exceed 0.2 m/s. The stress of heat loss and chilling has been found to negatively affect the calf’s immune system, leading to a higher susceptibility of different pathogens (Jennings and Glover, 1952). In our study in herds whose calf pens were placed along an outer wall, the calves had an increased risk of suffering from diarrhoea (OR: 1.92) and tended also to have an increased risk of other infectious disease (P < 0.10). The association between the placing of the calf pens and disease could not be explained by draught. Alternative explanations for the effect may be related to cold radiation, high relative humidity and damp bedding conditions. Calves in herds where antibodies against bovine virus diarrhoea virus or persistently infected animals were present at the start of the study ran an increased risk of having moderately to severely increased respiratory sounds at lung auscultation compared to herds free of the infection. Larsson et al. (1994) found calf mortality and the percentage of calves treated by veterinarian for respiratory disease and/or enteritis to be significantly higher in calves born during the period of BVDV introduction to the herd than in calves born the previous year or the year after (when calves persistently infected with the virus was present for most of the period). Similar findings had previously been reported by Barber et al. (1985). The primary infection with BVDV has been shown to alter the function of neutrophils and depress the proliferative rate of lymphocytes when tested in vitro (Roth and Kaeberle, 1983). Tra˚ ve´ n et al. (1991) found a transient leukopenia and lymphopenia in the acute phase of the infection. Therefore, the virus appears to be an immunosuppressive agent which may increase the susceptibility of the host to other respiratory or enteric pathogens (Potgieter et al., 1984; Wray and Roeder, 1987). It may also potentate or exacerbate the pathogenicity of co-infecting pathogens and may thereby increase the severity of concomitant diseases (Potgieter, 1988; de Verdier Klingenberg, 2000). The risk for respiratory disease and increased respiratory sounds was increased on farms with poor record-keeping while the opposite was found for diarrhoea. With regard to diarrhoea, farmers who had a ‘‘less good’’ record-keeping capacity probably did not notice all the cases of diarrhoea. A similar assumption between increased milk production at the
140
G.K. Lundborg et al. / Preventive Veterinary Medicine 68 (2005) 123–143
herd level and increased disease treatment frequencies has been made by Emanuelson and Oltenacu (1998). Respiratory disease was largely diagnosed by the project veterinarians who also performed all the lung-auscultations. The farmers’ thoroughness in keeping records of calf diseases is, however, likely to be related to their thoroughness in managing the calves, which could of course have an impact on the disease risk. We did not find the sex and status of the caretaker to be associated with the herd-level morbidity. The literature reports conflicting results concerning this. Speicher and Hepp (1973) reported higher calf losses when hired staff rather than family members cared for the calves. Whereas Hartman et al. (1974), Martin et al. (1975) and Jenney et al. (1981) reported similar results, Oxender et al. (1973), James et al. (1984) and Hagstad et al. (1984) could not demonstrate statistically significant differences between animals cared for by hired staff and animals tended by the owner. James et al. (1984) found a lower mortality when the wife or children cared for the calves. Trends of better results for the owner’s mother or wife were also reported by Hartman et al. (1974) and Speicher and Hepp (1973). Waltner-Toews et al. (1986a) reported that during the summer, family members other than owner had less calf mortality than the owner him/herself and discussed how factors such as methods of grouping and a small number of farms with hired staff could affect the results. Also in the present study, only a small number of farms had hired staff. However, the fact that the capacity of keeping accurate records was found to have an effect on the disease risks may be viewed as an indicator of the caretakers’ importance. It is well known that sudden changes in the diet may result in diarrhoea. Wray and Thomlinson (1975) observed that increases in the proportion of total pathogenic Escherichia coli in the faeces of young calves often followed changes in the liquid diet. We did not find an association between diarrhoea and method of weaning. This could be due to the fact that farmers may have failed to register cases of diarrhoea at weaning, either because the period of faeces looser than normal was shorter than our stipulated 2 days or because they considered diarrhoea at weaning as normal. In the present study, no significant association was found between the cleanliness of the bedding and calf morbidity. Curtis et al. (1993) found damp bedding to be a risk factor for enteritis within 14 days after birth.
5. Conclusion The result of this study stresses the impact of environmental and management factors and gives more evidence for the importance of an infection with BVDV. The advices that could be extrapolated from this study are that calf pens should be placed separated from outer walls and kept draught-free and that keeping the herd BVDV-free is a good investment in the health of the calves.
Acknowledgements This study was financially supported by the Swedish Farmers’ Foundation for Agricultural Research, the Swedish Council for Forestry and Agricultural Research, the
G.K. Lundborg et al. / Preventive Veterinary Medicine 68 (2005) 123–143
141
Sko¨ vde and Va¨ xjo¨ Djurskyddsfo¨ reningar, Stiftelsen Skaraborgs No¨ tkreaturfo¨ rsa¨ kringsbolagsfond, Agrova¨ st, the Swedish Dairy Association and the Bro¨ derna Johnssons Forskningsfond. We would like to thank the participating farmers. We would also like to extend thanks to Kerstin Plym Forshell (formerly of Svensk Mjo¨ lk; present affiliation: Tine Norske Mejerier), Sven-Ove Olsson (of Svensk Mjo¨ lk), and Ulf Emanuelson (of Epi-Lys) for their help in designing the study, and Jan Nilsson (of the Swedish University of Agricultural Sciences) for help with the editing of data. We are also grateful to Lotta Andersson and ¨ stlund (of the Swedish University of Agricultural Sciences) and the staff at Skara Jonica O Semin for their help with collection of data, and to Gunilla Jacobsson (of the Swedish University of Agricultural Sciences) for practical data handling. Finally, we would like to thank Daniel Maizon (of the Cornell University) and Jan Hultgren (of the Swedish University of Agricultural Sciences) for valuable help with the statistical analyses.
References Alenius, S., Lindberg, A., Larsson, B., 1997. A national approach to the control of bovine viral diarrhoea virus. In: Edward, S., Paton, D.J., Wensvoort, G. (Eds.), Third ESVV Symposium on Pestivirus Infections, 19–20 September 1996, Lelystadt, The Netherlands, pp. 162–169. Andrews, A.H., 2000. Calf pneumonia costs! BVCA 8, 109–114. Barber, D.M.L., Nettleton, P.F., Herring, J.A., 1985. Disease in a dairy herd associated with the introduction and spread of bovine virus diarrhoea virus. Vet. Rec. 117, 459–464. Curtis, C.R., Erb, H.N., Scarlett, J.M., White, M.E., 1993. Path model of herd-level risk factors for calfhood morbidity and mortality in New York Holstein herds. Prev. Vet. Med. 16, 223–237. Curtis, C.R., Scarlett, J.M., Erb, H.N., White, M.E., 1988. Path model of individual-calf risk factors for calfhood morbidity and mortality in New York Holstein herds. Prev. Vet. Med. 6, 43–62. de Verdier Klingenberg, K., 2000. Enhancement of clinical signs in experimental rotavirus infected calves by combined viral infections. Vet. Rec. 147, 717–719. de Verdier Klingenberg, K., Va˚ gsholm, I., Alenius, S., 1999. Incidence of diarrhea among calves after strict closure and eradication of bovine viral diarrhea virus infection in a dairy herd. J. Am. Vet. Med. Assoc. 214, 1824– 1828. Emanuelson, U., Oltenacu, P.A., 1998. Incidences and effects of diseases on the performance of Swedish dairy herds stratified by production. J. Dairy Sci. 81, 2376–2382. Ersbøll, A.K., Rugbjerg, H., Stryhn, H., 2001. Increased mortality among calves in Danish cattle herds during bovine virus diarrhoea infection.. XIth International Conference on Production Diseases in Farm Animals, Fredriksberg, Denmark, August 12–16, pp. 77–78. Fourichon, C., Beaudeau, F., Seegers, H., 1997. Critical points related to housing and management in control programmes for calf morbidity and mortality in French dairy herds.. In: Proceedings of the Ninth International Congress on Animal Hygiene, Helsinki, Finland, August 17–21, pp. 32–35. Goldstein, H., 1995. Multilevel Statistical Models. Edward Arnold/Wiley, London/New York. Goldstein, H., Browne, W.J., Rasbach, J., 2002. Partitioning variation in multilevel models. Understanding Stat. 1, 223–241. Goldstein, H., Rasbach, J., 1996. Improved approximations for multilevel models with binary responses. J. R. Stat. Soc. 159, 505–513. Gunn, G.J., Stott, A.W., 1997. A comparison of economic losses due to calf enteritis and calf pneumonia in northern Scotland. In: Proceedings of the VIIIth International Symposium of Veterinary Epidemiology and Economics, Paris, France, July 8–11, pp. 10.06.1–10.06.3. Hagstad, H.V., Nicholson, S.S., Fulton, R.W., Springer, W.T., Cox, H.U., 1984. Influence of management practices on dairy calf mortality. Trop. Vet. 2, 123–127.
142
G.K. Lundborg et al. / Preventive Veterinary Medicine 68 (2005) 123–143
Harry, E.G., 1978. Air pollution in farm buildings and methods of control: a review. Avian Pathol. 7, 441–454. Hartman, D.A., Everett, R.W., Slack, S.T., Warner, R.G., 1974. Calf mortality. J. Dairy Sci. 57, 576–578. Houe, H., Meyling, A., 1991. Surveillance of cattle herds for bovine virus diarrhoea virus (BVDV)-infection using data on reproduction and calf mortality. Arch. Virol. Suppl. 3, 157–164. James, R.E., McGilliard, M.L., Hartman, D.A., 1984. Calf mortality of Virginia dairy herd improvement herds. J. Dairy Sci. 67, 908–911. Jenney, B.F., Gramling, G.E., Glaze, T.M., 1981. Management factors associated with calf mortality in South Carolina dairy herds. J. Dairy Sci. 64, 2284–2289. Jennings, A.R., Glover, R.E., 1952. Enzootic pneumonia in calves. II. The experimental disease. J. Comp. Pathol. 62, 6–22. Kilburn, K.H., 1967. Cilia and mucus transport as determinants of the response of the lung to air pollutants. Arch. Environ. Health 14, 77–90. Kiorpes, A., Butler, D., Dubielzieg, R., Beck, K., 1988. Enzootic pneumonia in calves: clinical and morphologic features. Comp. Cont. Educ. Pract. Vet. 10, 248–258. Kirkwood, J.K., Hubrecht, R., 2001. Animal consciousness, cognition and welfare. Anim. Welf. 10, 5–17. Larsson, B., Niskanen, R., Alenius, S., 1994. Natural infection with bovine virus diarrhoea virus in a dairy herd: a spectrum of symptoms including early reproductive failure and retained placenta. Anim. Reprod. Sci. 36, 37– 48. Martig, J., Boss, P.H., Nicolet, J., Steck, F., 1976. Etiology and predisposing factors in respiratory disease of milkfattened veal calves. Livestock Prod. Sci. 3, 285–295. Martin, S.W., Schwabe, C.W., Franti, C.E., 1975. Dairy calf mortality rate: influence of management and housing factors on calf mortality rate in Tulare County, California. Am. J. Vet. Res. 36, 1111–1114. MLwiN Help Version 3.0.02, March 2002. # Centre for Multilevel Modelling, Institute of Education. Moerman, A., Straver, P.J., de Jong, M.C., Quak, J., Baanvinger, T., van Oirshot, J.T., 1994. Clinical consequences of a bovine virus diarrhoea virus infection in a dairy herd: a longitudinal study. Vet. Q. 16, 115–119. Olsson, S.-O., Jacobsson, L., Alenius, S., Larsson, B., 1993. A voluntary control programme against infection with bovine diarrhoea virus (BVDV). Svensk Vet. Tidn. 45, 411–415. Oxender, W.D., Newman, L.E., Morrow, D.A., 1973. Factors influencing dairy calf mortality in Michigan. J. Am. Vet. Assoc. 162, 458–460. Pettersson, K., Svensson, C., Liberg, P., 2001. Housing, feeding and management of calves and replacement heifers in Swedish dairy herds. Acta Vet. Scand. 42, 465–478. Potgieter, L.N.D., 1988. Immunosuppression in cattle as a result of bovine viral diarrhea virus infection. Agric. Pract. 9, 7–14. Potgieter, L.N.D., McCracken, M.D., Hopkins, F.M., Walker, R.D., 1984. Effect of bovine viral diarrhea virus infection on the distribution of infectious bovine rhinotracheitis virus in calves. J. Am. Vet. Res. 45, 687–690. Rasbash, J., Browne, W., Goldstein, H., Yang, M., Plewis, I., Healy, M., Woodhouse, G., Draper, D., Langford, I., Lewis, T., 2002. A User’s Guide to MLwiN. Version 2.1d for Use with MLwiN 1.10. Centre for Multilevel Modelling, Institute of Education, London. Roth, J.A., Kaeberle, J.W., 1983. Suppression of neutrophil and lymphocyte function induced by a vaccinal strain of bovine viral diarrhea virus with and without the administration of ACTH. Am. J. Vet. Res. 44, 2366–2372. Simensen, E., Norheim, K., 1983. An epidemiological study of calf health and performance in Norwegian dairy herds. IV. Factors affecting morbidity and performance. Acta Agric. Scand. 33, 65–74. Speicher, J.A., Hepp, R.E., 1973. Factors associated with calf mortality in Michigan dairy herds. J. Am. Vet. Med. Assoc. 162, 463–466. Svensson, C., Lundborg, K., Emanuelson, U., Olsson, S.-O., 2003. Morbidity in Swedish dairy calves from birth to 90 days of age and individual calf-level risk factors for infectious diseases. Prev. Vet. Med. 58, 179–197. Tra˚ ve´ n, M., Alenius, S., Fossum, C., Larsson, B., 1991. Primary bovine viral diarrhoea virus infection in calves following direct contact with a persistently viraemic calf. J. Vet. Med. B 38, 453–462. Virtala, A.M.K., Gro¨ hn, Y.T., Mechor, G.D., Erb, H.N., 1999. The effect of maternally derived immunoglobulin G on the risk of respiratory disease in heifers during the first 3 months of life. Prev. Vet. Med. 39, 25–37. Waltner-Toews, D., Martin, S.W., Meek, A.H., 1986a. Dairy calf management, morbidity and mortality in Ontario Holstein herds. III. Association of management with morbidity. Prev. Vet. Med. 4, 137–158. Waltner-Toews, D., Martin, S.W., Meek, A.H., McMillian, I., 1986b. Dairy calf management, morbidity and mortality in Ontario Holstein herds. I. The data. Prev. Vet. Med. 4, 103–124.
G.K. Lundborg et al. / Preventive Veterinary Medicine 68 (2005) 123–143
143
Wathes, C.M., 1998. Environmental control in pig housing. In: Done, S., Thomson, J., Varley, M. (Eds.), XVth International Pig Veterinary Society Congress, Birmingham. Nottingham University Press, Nottingham, Great Britain, pp. 257–265. Wathes, C.M., Jones, C.D.R., Webster, A.J.F., 1983. Ventilation, air hygiene and animal health. Vet. Rec. 113, 554–559. Wray, C., Roeder, P.L., 1987. Effect of bovine virus diarrhoea–mucosal disease virus infection on salmonella infection in calves. Res. Vet. Sci. 42, 213–218. Wray, C., Thomlinson, J.R., 1975. Factors influencing the occurrence of colibacillosis in calves. Vet. Rec. 96, 52–56.