Материал: Bovine Viral Diarrhea Virus Diagnosis, Management, and Control

etiology and risk factor is unclear. However, this is more of an academic discussion.

The importance of different risk factors will vary substantially between regions. For example, the purchase of animals without testing will be a much higher risk factor in high-prevalence areas such as central Europe than in low-prevalence areas such as Finland. The following examples on risk factors should therefore be seen in the context of the prevalence, as shown in Tables 2.1–2.3.

In a cross-sectional study, the prevalence of antibody carriers was significantly higher among animals kept in extensive production systems (76.8%) than those in intensive productions systems (70.9%) (Reinhardt et al., 1990). This was believed in part to be due to a higher rate of animal movement in extensive production systems. In the same study there was no difference in prevalence between beef, dairy, or mixed farming.

A case-control study of 314 herds in Norway compared risk factors between herds with antibodypositive young stock (case herds) and those with negative bulk tank milk samples (control herds) (Valle et al., 1999). A high number of risk factors occurred in a higher percentage of case herds as compared to control herds, including (percentage of case herds compared to control herds given in parenthesis): heifers on common pasture (21.0 vs. 8.7), sheep in pasture with cattle (31.5 vs. 22.1), breaking through pasture fences (33.8 vs. 24.6), over pasture fence contact (50.0 vs. 31.2), mixing of herds in pasture (43.8 vs. 30.6), wild animals in pasture (92.2 vs. 84.5), purchasing animals (67.3 vs. 48.7), not asking information about BVD when purchasing animals (73.4 vs. 40.9), other animal traffic including common summer housing and exchange of calves (9.6 vs. 0.6), veterinarian reusing needles between farms (7.1 vs. 1.3), and not using dairy advisors (25.0 vs. 12.0). However, due to the long list of variables and occasional low number of herds with the indicated risk factor, few risk factors were significant. In the logistic regression models, purchasing animals (OR = 1.8), use of common pasture (OR = 5.1), over pasture fence contact (OR = 2.5), purchasing without BVD documentation (OR = 5.4), not using dairy advisors (OR = 4.1), being a younger farmer, and “other animal traffic” (OR = 28.6) were found to be risk factors, which explained 51% of the seropositive herds.

In a repeat cross-sectional study in Denmark, 41 dairy herds were classified into lightly infected ( 3 antibody carriers among 10 young stock) and highly infected (≥8 antibody carriers among 10 young

stock) in 1992 and 1994 (Houe et al., 1997; Houe, 1999). Of the 24 herds being lightly infected in 1992, 11 changed to highly infected in 1994 and 13 remained lightly infected. The change in infection status was associated with the purchase of new animals within the past 3 years (P = 0.052) and moderately associated with pasturing cattle at a distance of less than 5 m to cattle from other herds or having other contacts with cattle from other herds (P = 0.085).

Risk factors for the presence of PI animals in cattle herds have been studied including demographic information on herd sizes and information from geographic information systems concerning distances between herds. Multivariable logistic regression with data from more than 8,000 herds showed significant effect of the following risk factors (OR, CI, and p values given in parenthesis) on the occurrence of PI animals: herd size measured as number of cows (OR = 1.09 per change in herd size of 10 cows; CI = 1.06–1.11; P<0.001), mean distance to neighboring herds (OR = 0.87 per change in unit of 500m; CI = 0.81–0.93; P<0.001) and sum of infected neighbors (OR = 1.54 for herds with >3 infected neighbors as compared to no infected neighbors; CI = 1.17–2.02; P = 0.024 for the overall effect of sum of infected neighbors) (Ersbøll and Stryhn, 2000). In a study in U.S. herds with PI animals were significantly (P<0.01) larger than herds without PI animals (Wittum et al., 2001). Thus, herds with PI animals had a medium number of 245 calves born in the calving season as compared to 89 in herds without PI animals. The OR for a herd of having high levels of antibodies in bulk tank milk was shown to increase by a factor 1.003 for each additional cow in the herd (Paton et al., 1998).

Several risk factors are confounded among each other. If, for example, larger herds purchase more animals than smaller herds then an apparent effect of herd size may in reality be due to purchase of animals and not to the herd size. It is, therefore, important that risk factor studies are performed by including several risk factors. Furthermore, there is a need to address interactions between risk factors. For example, the risk of infection at pasture will be higher in a high-prevalence area as compared to a lowprevalence area.

An alternative to large-scale studies identifying risk factors for the presence of infection is to perform intensive follow-up in recently infected herds. Follow-up investigations were performed in 67 previously BVDV-free Danish herds that got infected in 1998. Obvious explanations for re-infection were

identified in 74% of the cases. Of the 67 herds, 28% had purchased pregnant animals that later delivered PI animals. In 36% of the herds, PI animals had been present on neighboring pastures. Seven percent of the herds had had animals on common pasture and in 3% of the herds there had been PI animals in neighboring farmhouses. In the remaining 26% of the herds, no obvious explanations could be identified (Bitsch et al., 2000). Because this is an explorative type of study, probabilities cannot be given for the actual reasons apart from the given frequency distribution of the most obviously identified reasons.

As previously stated, the effect of risk factors will be highly influenced by the prevalence of infection in the area. Therefore the effect of different risk factors will obviously change during an eradication phase. It can be beneficial to combine different studies in order to obtain sufficient observational units for statistical analysis. Attempts have been made to compare risk factors in a low-prevalence area (Michigan) with that in a high-prevalence area (Denmark) (Houe et al., 1995a). The actual prevalence related to two previous studies (Houe and Meyling, 1991; Houe et al., 1995b; prevalence can be seen in Tables 2.1 and 2.2) showed that the prevalence of PI animals was 10 times higher in Denmark than in Michigan. When the effect of purchasing any animal versus no purchase of animals was analyzed separately in the two areas there was no significant effect of purchase on the occurrence of PI animals in individual herds. However, when including both areas in the analyses stratifying on area, a significant effect of purchasing animals was found. The analyses also included a number of demographic variables (either obtained at the country or state level or obtained for the studied herds), and all risk factors except herd size were in favor of the lower prevalence in Michigan. The following examples where the variables for Denmark versus Michigan are given in parentheses illustrate the differences: concentration of cattle in the country or state (DK: 67 per km2, MI: 27 per km2), concentration of herds in the country or state (DK: 1 per km2, MI: 0.29 per km2), herd size among the studied herds (DK: 135, MI: 274), use of pastures (DK: 79% of herds, MI: 45% of herds), purchase and addition of animals to the studied herds (DK: 18.5% of herds; MI: 12.6% of herds), and use of vaccination (DK: no vaccination, MI: 75% of studied herds). On a regional basis a study in England and Wales showed a significant relationship between antibody level in bulk tank milk and cattle population density (Paton et al., 1998).

With the high number of prevalence studies performed in recent years it would be an obvious next step to gather all the information on prevalence studies with information on cattle demographics (e.g., cattle density, herd density, herd size, trade patterns, use of vaccination) from a number of regions and perform formal meta-analyses of possible risk factors. Already some trends are obvious (Tables 2.1–2.3). For example, low-prevalence areas such as Norway and Finland have lower cattle population density and smaller herd size than high-prevalence areas (or areas that used to be high-prevalence areas before an eradication program), such as Belgium, Germany, and Denmark.

Despite the intuitively obvious effects of many of these risk factors, the documentation in literature is often scarce and the risk factors can explain only a relatively low percentage of infections. Some of the reasons for difficulties in establishing the risk factors may be due to uncertainties of the time of infection. For example, presence of infection in a herd may be due to introduction of infection 3–4 years ago as illustrated in the following study. Routinely recorded data (register data) from the Danish Cattle Database and the BVDV eradication scheme were used to estimate the origin of the first PI animal occurring in Danish cattle herds (Alban et al., 2001). The study showed that among herds participating in the milk recording scheme (i.e., dairy herds) the first PI animal was born to a dam that had been in the herd the whole period of lactation in 76.1% of the cases, PI animals were introduced directly in 4.5% of the cases and they were introduced by purchase of a dam in late gestation in 4.3% of the cases. In the remaining cases the exact source could not be identified. For comparison, among herds not in the milk recording scheme (mostly beef herds), the first PI animal was born from a dam that had been in the herd during the whole lactation in 26.3% of the cases, whereas PI animals had been directly introduced in 33.9% of the cases and through dams in late lactation in 2.4% of the cases. It should be noted that some of the dams that had been in the herds throughout their whole lactation could have been infected by previously purchased but unidentified PI animals. Still, most of the PI animals in the herds were generated within the herd, meaning that the exact trace back for risk factors often needs to go several years back in time. Therefore, cohort studies linking the time of infection with the time of occurrence of risk factors would be useful to further quantify the importance of risk factors.

QUANTIFICATION OF THE EFFECTS/CONSEQUENCES OF BVDV INFECTIONS ON DISEASE AND PRODUCTION

The measures of consequences include the presence of the disease as well as reduced production. The clinical manifestations often used in epidemiological studies to measure the effect of transient infection include diarrhea, reduced milk production, reproductive disorders, increased occurrence of other diseases, and death. Losses from fetal infection include abortions, congenital defects, birth of weak and undersized calves, unthrifty PI animals, and death among PI animals (Baker, 1995). These consequences of BVDV infections have been documented in several experimental and case studies, whereas quantification of the consequences in larger observational studies is often less well documented. The validity of these data in relation to their verification of being BVD-related varies substantially. The losses following acute (transient) infection are especially difficult to quantify. Losses such as respiratory disease, repeat breeding, or abortions have seldom been measured together with a rise in antibody titer, and therefore such losses are often analyzed in retrospective studies, where the time period of transient infections is only loosely estimated. Losses among the PI animals have been quantified with a higher degree of certainty because the presence of persistent infection together with clinical symptoms has often been documented. Measures of prevalence and incidence together with quantification of disease and production parameters are the foundation for calculating the effect of BVDV in economical terms. The following sections emphasize measures of consequences that have been quantified under field conditions followed by estimation of economic impact of BVDV infections. On some occasions quantitative assessments from case studies and experimental studies are also given when they are judged to have sufficient quantitative interpretation.

The effects of infection have often been investigated in case studies and experimental studies, but to a lesser extent in observational studies. All of these approaches have drawbacks associated with them. For example, more severe outbreaks tend to attract attention and are liable to be investigated and published leading to overestimation of the effect of BVDV infections. In experimental studies, it is difficult to mimic natural circumstances and it has often been difficult to demonstrate clinical signs of acute BVD, although reproductive disorders (e.g.,

congenital defects) have frequently been documented. Observational studies often suffer from the lack of knowledge of how many animals actually seroconverted and at what time points. Therefore, the effects seen in observational studies are probably underestimated when considering the effect in the individual seroconverting animal. However, observational studies are useful for estimation of the effect of infection in a partly immune population.

SUBCLINICAL INFECTIONS

The clinical spectrum of postnatal infection of immunocompetent animals ranges from subclinical infection to severe disease followed by death. Several years ago it was estimated that 70–90% of transient infections were mild or subclinical (Ames et al., 1986). Although the risk of disease following transient infection has not been exactly quantified, epidemiological investigations indicate that these estimates are good. The clinical consequences of acute BVDV infection were studied in a large dairy herd where BVDV was circulating for approximately 2.5 years (Moerman et al., 1994). Among 136 cattle with transient infection, 129 (95%) showed no clinical signs of infection, 5 animals were slightly affected, and only 2 animals were severely ill. Furthermore, many studies in the 1990s have shown that BVDV infection was more widespread than previously thought, indicating the presence of a large number of subclinical infections.

REDUCED MILK YIELD

Milk yield was studied in a longitudinal study of 54 cows of which 22 seroconverted (Moermann et al., 1994). A gradual drop in the milk yield (measured as moving average over 3 days) of 10% or more within 10 days was seen in 18 out of 22 cows that seroconverted and 9 out of 32 cows that did not seroconvert, giving an OR of reduced milk production of 11.5. On the other hand comparison of antibody level in bulk tank milk and milk yield did not reveal a significant association (Graham et al., 2001).

OCCURRENCE OF OTHER DISEASES

FOLLOWING TRANSIENT INFECTION IN

COWS

Examination of bulk tank milk samples from 213 Swedish dairy herds in 1988 and 1989 indicated that recent infection had occurred in 7 herds (Niskanen et al., 1995). Compared to 84 herds with continuous low level of antibodies in bulk tank milk (i.e., uninfected), cows in the 7 infected herds had increased ORs for several clinical conditions, with mastitis,

miscellaneous diseases, and retained placenta being significant (OR and 95% CI for OR are given in parenthesis): mastitis (OR = 1.8; CI = 1.1–2.8), tramped teat (OR = 1.6; CI = 0.9–2.8), ketosis (OR = 1.3; CI = 0.6–2.2), miscellaneous diseases (OR = 2.8; CI = 1.7–4.4) and retained placenta (OR = 2.8; CI = 1.6–4.7). Increased risk of retained placenta may also be seen among dams delivering PI calves (i.e., several months after the transient infection). Thus, 5 (41.7%) out of 12 PI dams had retained placenta compared to 7 (3.5%) cases of retained placenta in 198 non-PI dams (P<0.001) (Larsson et al., 1994).

In Norway, the national screening of all dairy herds for BVDV antibodies in bulk milk in 1992 and 1993 was used to estimate the effect of transient infection on udder health (Waage, 2000). A total of 404 herds with a significant rise in antibody level in bulk milk were compared to herds with continuous low levels of antibodies. In infected herds, there was a 7% (CI = 0.2%–11.4%) increase in the incidence of clinical mastitis in the year of exposure as compared to negative control herds but there was no significant effect on the bulk milk somatic cell count or the culling rate due to mastitis. However, a study by Graham et al. (2001) found a significant increase in somatic cell count with increasing antibody levels in bulk tank milk (P<0.01).

OCCURRENCE OF OTHER DISEASES

FOLLOWING TRANSIENT INFECTION IN

CALVES

Transient BVDV infections have been associated with respiratory diseases and diarrhea in calves. However, due to the multifactorial nature of these diseases it is difficult to quantify the exact role of BVDV in their occurrence. Different field investigations indicate that BVDV will often double the risk of these conditions. Among calves entering a feedlot, 13 of 29 (45%) calves treated for respiratory disease and 8 of 36 (22%) untreated calves had seroconverted (Martin and Bohac, 1986). This corresponds to an OR of 2.8 for treatment following seroconversion. The effect of BVDV infection on the occurrence of respiratory diseases has also been demonstrated indirectly by comparing calves receiving colostrum from seronegative dams with calves receiving colostrum from seropositive dams (Moerman et al., 1994). Among calves exposed to BVDV soon after birth, 30 of 44 calves (68.2%) receiving colostrum from seronegative dams developed moderate or severe bronchopneumonia whereas 35 of 86 calves (40.7%) receiving colostrum from seroposi-

tive dams developed moderate or severe bronchopneumonia. This corresponds to an OR of 3.1 for developing bronchopneumonia if receiving colostrum from antibody-negative dams. In another herd 10 of 61 calves (16.4%) born in the period when BVDV was introduced into the herd, underwent veterinary treatment for respiratory disease and/or enteritis between 2 days and 4 months of age, whereas only 5 out of 134 calves (3.7%) born before or after the introduction period received treatment (P<0.01) (Larsson et al., 1994). This corresponds to an OR of 5.1 of being treated if born in the period of BVDV introduction. In the same study the mortality was 13.1% among calves born in the period of introduction of BVDV as compared to 2.2% in the other periods (P<0.01).

After strict closure of a dairy herd together with eradication of BVDV the incidence of diarrhea among calves in the first 31 days of life decreased significantly from 70.6% to 19.4% (Klingenberg et al., 1999). However, it was difficult to distinguish the strict effect of BVDV compared to general improvement in management practice including closure of the herd.

In addition to having an effect on clinical diseases, BVDV also has an effect on the occurrence of other infections. For example infectious bovine rhinotracheitis (IBR) virus was much more widespread in the respiratory tract of calves previously exposed to BVDV (Potgieter et al., 1984a). Furthermore, there are many examples that BVDV will aggravate the severeness of other infections—for example, P. haemolytica (Potgieter et al., 1984b; also see Chapter 9 on immunity and immunosuppression). However, the quantitative effect of BVDV on the prevalence, incidence, and severity of other infections is not completely documented.

TRANSIENT INFECTION BY VIRULENT

STRAINS OR BVDV IN COMBINATION WITH

OTHER PATHOGENS

For many years postnatal infection of immunocompetent animals was characterized as mild or subclinical in most cases. In the 1990s there came increased evidence of the importance of highly virulent strains with much more severe consequences following transient infection (hemorrhagic syndrome and peracute BVD) especially with BVDV genotype 2. Although there is no unambiguous connection between genotype and virulence, genotype 2 has predominated as the cause of hemorrhagic syndrome and peracute BVDV (Bolin and Ridpath, 1996; Ridpath et al., 2000).

An outbreak of acute BVD in a herd also infected with Leptospira hardjo and Coxiella burnetii caused very severe losses (Pritchard et al., 1989). In a dairy herd with 183 milking cows, the losses included 15 dead cows, 20 culled cows (primarily because of emaciation and infertility), 40 abortions, 18 stillborn calves, and 3 calves dying soon after birth. Other case descriptions of acute outbreaks have revealed morbidity of up to 40% and mortality up to 10% (Hibberd et al., 1993; David et al., 1994).

In 1993, BVDV genotype 2 was associated with severe outbreaks of BVDV in Ontario, Canada. A study of seven herds (mean number of cattle: 93; range: 40–191) with outbreaks of acute BVD revealed a mortality among adults of 9% (range 2–26%), a mortality among immature animals (<2 years of age) of 53% (range 13–100%), and 44% (range 3–83%) mean abortion frequency based on the number of breeding age females. Furthermore, the occurrence of respiratory diseases was high along with a number of sequelae (Carman et al., 1998). In 1993, the mortality was estimated as 31.5% for grain-fed calves and 17.1% for milk-fed calves, and the overall mortality was 32,000 out of 143,000 calves (Pellerin et al., 1994).

REPRODUCTIVE DISORDERS

Reproductive disorders are a very important clinical manifestation. These are outlined in more detail in Chapter 8. This chapter emphasizes the quantitative aspects of their occurrence. An often-used measure for reproduction efficiency is the conception rate. Note in the following examples that although the measure is usually defined as a proportion it is still termed conception rate (although it is mathematically not a rate). In an experimental study, 15 heifers (Group I) were infected by contact with PI animals 4 days after insemination, another 18 heifers (Group II) were infected intranasally 9 days before insemination, and 14 uninfected heifers acted as control group (McGowan et al., 1993). The conception rates 20 days after insemination (defined as proportion of eligible heifers that had serum progesterone concentration of more than 2.0 ng/ml on day 20 after insemination) were 60% (9 of 15) and 44% (8 of 18) for Groups I and II as compared to 79% (11 of 14) in the control group. At 77 days after insemination the conception rates (defined as proportion of eligible heifers which were palpably pregnant 77 days after insemination) were 33% (5 of 15) and 39% (7 of 18) for the infected groups; those for the control group remained unchanged. At 77 days, conception rates in Groups I and II were both significantly

lower than the control group. A group of seronegative cows that were inadvertently exposed to a PI animal were bred before (9 cows), during (9 cows), or after seroconversion (14 cows) (Virakul et al., 1988). The first service conception rates at 21 days in these three groups were 22.2%, 44.4%, and 78.6% (the difference between 22.2% and 78.6% being significant at P<0.05).

The severe effect on conception rates is also seen in observational studies, although the effect is somewhat less pronounced. In five Danish dairy herds, the overall conception rate in a risk period of infection (presence of a young PI animal spreading the infection) was 38% as compared to a conception rate of 47% (P<0.001) in a post-risk period (presence of older PI animals where the herd would be immune) (Houe et al., 1993a). The increase in conception rate ranged from 16–64% in individual herds.

In seven Swedish herds with recent infection there was an increased OR of estrus stimulating treatments (OR = 2.2; CI = 1.0–4.9); however, the number of inseminations per service period was not significantly increased (Niskanen et al., 1995). In Norway the ORs for abortions in herds with recent infection was calculated as 2.6 and 11.6, respectively for two different registration periods (P< 0.01) as compared to BVDV free control herds (Frederiksen et al., 1998). A Swiss study showed that animals seroconverting in midgestation (days 46–210) had an OR of abortion of 3.1 (P = 0.02), whereas an effect was not seen after infection in later stages of gestation (Rüfenacht et al., 2001). In the study by Moerman et al. (1994), abortions were not more frequent in seroconverting cattle. A field investigation of the causes of abortions in the United Kingdom showed that 39 out of 149 (26%) abortions in 54 farms were associated with BVDV (Murray, 1990).

Although congenital defects, stillbirths and weakborn calves have been demonstrated in many experimental studies (e.g., Done et al., 1980; Liess et al., 1984; McClurkin et al., 1984; Roeder et al., 1986), they have seldomly been quantified in larger observational studies. Some observational studies did not find significantly more stillbirths, weak-born calves, or congenital defects (Moerman et al., 1994; Frederiksen et al., 1998; Rüfenacht et al., 2001).

LOSSES AMONG PI ANIMALS

PI animals are often born weak and undersized and when they grow up they are at risk of acquiring other diseases and being culled due to unthriftiness or they may die from mucosal disease. In a study the