The epidemiology of bovine respiratory disease: What is the evidence for preventive measures?

Abstract

Bovine respiratory disease (BRD) is the most common and costly disease of beef cattle in North America. Despite extensive research, industry practices are often more informed by dogma than by fact. Frequently advocated interventions, including vaccination, various processing procedures, and nutritional manipulation, have limited impact on morbidity and mortality. Evidence for use of oral antimicrobials, either in feed or water, appears to be equivocal. In contrast, preconditioning and metaphylaxis have significant scientific evidence of efficacy, with weaning prior to sale potentially being the most important component of preconditioning. The inability to reach more definitive conclusions in preventing BRD may be attributable to difficulties in investigating the disease. Study challenges include potential for extensive confounding, tremendous variability, the multi-factorial nature of the disease, and inadequate methods for diagnosis.

Introduction

Bovine respiratory disease (BRD) is the most costly disease of beef cattle in North America (1). It is also one of the most extensively studied, with research beginning in the late 1800s and continuing today. A recent review examined various proposed predisposing factors (2). This article will examine evidence regarding efficacy of efforts to reduce severity and frequency of BRD. Commentary is also made on the challenges that complicate investigation of natural BRD.

Interventions

Processing

Cattle arriving at feedlots or stocker operations typically are subjected to a series of activities collectively known as processing. These procedures include administration of vaccines, growth promotant implants, anthelmintics and parenteral vitamins, as well as castration, dehorning, and abortion of heifers found to be pregnant. Unfortunately, these activities have received little research in relation to BRD. The effects of castration and dehorning on BRD were discussed recently (2), and will not be reviewed here. The metabolic effects of growth-promotant implants are well known. Phillips et al (3) hypothesized that improved protein efficiency induced by zeranol would be beneficial for stressed calves. The study concluded that implants conserved energy reserves during transit, but morbidity was not affected. Martin et al (4) found that deworming in the first 2 wk after arrival had a negative impact on health. However, this association may have simply been co-linear with other processing procedures, including vaccination, which were associated with increased BRD. Another study found pre-conditioned calves dewormed with thiabendazole had lower disease incidence compared with calves not preconditioned and dewormed (5). But again, the possibility exists for co-linearity with other more important factors. Two studies found no difference in BRD morbidity in calves treated with an anthelmintic compared with those not treated (6,7). Comparison of anthelmintic products found no difference in BRD incidence among cattle treated with various products (8,9) or detected modest differences deemed economically unimportant (10).

The relationship of parenteral administration of vitamins A, C, D, and E with BRD was examined in an Australian study (11). Administration of vitamins A, D, and E at processing had no impact on BRD parameters; vitamin C, administered at the time of treatment, reduced mortality. In another study, vitamin E and selenium administration had no effect on BRD incidence, although antibody production in response to vaccination was enhanced (12).

Only 1 study examined the relationship between pregnancy evaluation and respiratory disease, with pregnancy examination being associated with increased morbidity (13). No attempt was made to explain this association. It could be speculated that heifers that were poorly managed prior to arrival were most likely to be examined for pregnancy, making this an indicator for other factors associated with increased risk for disease. Heifers that are found to be pregnant in the feedlot are typically aborted, and it has been hypothesized that abortion would be a stressful event that may contribute to BRD (14). While recommendations have been made to manage such heifers to reduce stress and disease (15), few studies have examined the relationship. One study examining open, aborted and pregnant heifers observed “no significant health problems or mortality in any of the groups,” but no data or statistical analysis was provided (16). A model designed to assess cost of various management options assumed a 5% increase in BRD morbidity associated with aborting heifers (17), but this figure was extrapolated from a study examining immunosuppression associated with dystocia rather than documented effects of abortion on respiratory disease.

Timing of processing

It has been theorized that added stress of processing procedures may contribute to morbidity. This has occasionally led to the recommendation to delay processing until cattle have acclimated to the new environment, or to perform some procedures at arrival while delaying others. Empirical evidence indicates that most in the industry do not consider processing at or shortly after arrival to be detrimental; a large percentage of feedlots employ this practice, and many of the studies reviewed (regardless of variables of interest examined) describe prompt processing of calves (13,18–26). One study reported a trend toward higher morbidity in calves when processing was delayed compared to processing at arrival. Delaying processing 2 to 3 wk also resulted in initial BRD cases occurring over an extended period of time (27). These trends were minor and significance cannot be assigned because no statistical analysis was performed. Kreikemeier et al (28) found no difference in morbidity or mortality, but improved performance associated with processing at arrival compared to waiting 3 wk. In contrast, Martin et al (4) found that processing calves in the first 2 wk after arrival increased disease compared with waiting 2 to 4 wk for processing. Others examined delaying some, but not all, components of processing. One study administered anti-parasitic treatments and clostridial vaccines to all cattle at arrival (as well as dehorning and castration). A modified live viral vaccine was either given at arrival, 1 wk after arrival, or not given at all (25). No difference attributable to timing of vaccination was noted; however, the power of this study to detect differences due to treatment was low. In another study, administration of a modified live viral vaccine was delayed 2 wk, whereas other processing activities (castration, deworming, and vaccination with a clostridial bacterin) were performed at arrival (29). Delaying viral vaccination resulted in increased average daily gain (ADG), with no difference in morbidity or mortality compared with inclusion at arrival.

Vaccination

Vaccination for viruses, and to a lesser extent bacterial pathogens, associated with BRD is widespread. Surprisingly, there is little scientific support for the practice. One of the first BRD epidemiologic studies found that vaccinating for respiratory pathogens in the feedlot was 1 of the most detrimental practices examined (4). In interpreting these data, it would appear feasible that calves considered at high risk for BRD were more likely to be vaccinated, thus creating a spurious association. This idea was disputed by the authors, who stated “Owners rarely vaccinate sick groups of calves and thus the increase of risk of death… is attributed to the vaccines” (30). Even accepting the authors’ interpretations, the data were collected from 1978 to 1980; thus it could be asserted that much has changed in vaccinology in the subsequent 3 decades. Indeed, no studies have replicated such significantly detrimental effects of vaccination. Nevertheless, when a rigorous critique is done of the entire body of work regarding respiratory vaccines, the evidence of efficacy is equivocal, at best (31,32). The presence of antibodies against various pathogens has been found to be protective (24,33), so it would seem intuitive that immunization would be beneficial. But while vaccination is consistently shown to result in antibody production, vaccine-induced titers are not always correlated with protection against disease (34). Much of the failure of vaccination may rest in timing of administration, failure of stressed calves to respond appropriately to vaccination, the multifactorial nature of BRD, and the increased susceptibility of stressed calves to all pathogens (obligate and opportunistic). These limitations can be somewhat overcome by utilizing vaccines prior to shipment, as part of a process termed preconditioning.

Preconditioning

Since no combination of receiving/processing practices has been found to eliminate BRD, significant effort has been extended in identifying pre-shipment practices that would reduce morbidity. Such measures, typically referred to as “preconditioning,” were proposed as early as 1967 (35). While variation exists in defining what constitutes “preconditioning,” common components are: vaccination for respiratory viruses and clostridial diseases (some programs also include M. haemolytica, P. multocida and H. somni); dehorning and castration, with adequate time allowed for healing prior to sale; weaning prior to sale, with time frames varying from several days to several weeks; training of calves to eat from a bunk and drink from a trough. The intent is to spread various stressors over multiple episodes rather than all at once. Additionally, vaccination at this time permits development of immunity prior to the time of maximum stress and exposure to pathogens.

Unfortunately, significant challenges arise in interpreting results of preconditioning studies. Many studies do not report on-farm morbidity occurring during the preconditioning phase. Thus it is possible that disease is not reduced, it is just shifted from the stocker or feedlot operator to the cow-calf producer. Frequently, preconditioned calves may benefit from being older and heavier at time of marketing, which may in and of itself reduce disease (2). Finally, some preconditioned calves are delivered directly to the feedlot or backgrounding operation, bypassing the commingling and stress associated with the sale barns from which control calves may be derived. Reduction in disease therefore cannot be ascribed solely to preconditioning. To truly assess the effect of preconditioning, control calves should also bypass the sale barns, but without weaning, vaccination and other requirements of preconditioning. Alternatively, preconditioned calves should be sold through the sale barn and order buyer process. Both of these approaches have been employed in recent studies, as discussed below.

Early peer-reviewed field trials evaluating the effectiveness of preconditioning were less than definitive in its benefit. Inconsistent results were obtained when multiple years were evaluated, such that morbidity was reduced in 1 year but not others (5). Some studies demonstrated no reduction of disease attributable to preconditioning (36), whereas others found reduction in disease but no economic advantage to preconditioning (37). A 1985 review verified the overall equivocal findings up to that point (38). On an aggregate basis, Cole (38) found a slight reduction in disease attributable to preconditioning, although the economic benefits were deemed questionable.

Many recent studies have demonstrated more positive results attributable to preconditioning (20,35,39–43). While these studies typically feature better design and analysis than previous research, most still have significant limitations. One major challenge in assessing these studies is lack of uniformity in defining preconditioning. For example, a study by Lynch et al (41) was more accurately a vaccination trial, as calves “were considered preconditioned if they had received both viral and Pasteurella vaccines 14 to 21 d prior to weaning.” Hansen et al (42) offered no description of what was required for calves to be considered pre-conditioned. Three other studies (20,39,40) required calves to be vaccinated and weaned to be called preconditioned. Macartney et al (40), however, had no requirement for “bunk breaking” or diet composition, whereas Roeber et al (20) and Karren et al (39) required calves to be acclimated to feed and water troughs. Such varied and vague definitions of “preconditioning” make across-study comparison difficult and results challenging to interpret. This potential confusion is further exemplified in a large survey that found no benefit to preconditioning (44). The preconditioning status of the calves was defined by the feedlot managers who reported the data; therefore, the definition was likely inconsistent, which may have influenced the findings.

As mentioned, how calves are marketed can affect the interpretation and relevance of preconditioning studies. In 2 studies (39,43), control and treatment calves came from 1 source and were sent directly to the feedlot. This eliminated much potential for variation but also reduced external validity, because most purchasers of preconditioned cattle acquire them from multiple sources and commingle them. Studies by Macartney et al (40) and Roeber et al (20) compared preconditioned calves to vaccinated calves and those sold through conventional sales (no vaccination or treatment requirements). Both preconditioned and vaccinated calves underwent vaccination, castration, and dehorning within defined periods prior to sale; in addition, calves in the preconditioned group were weaned at least 1 mo prior to sale. In each study, calves from the 3 groups (preconditioned, vaccinated without weaning and control) were marketed through the same sale barn (on different dates). These protocols are most consistent with industry definitions and practices, and make these studies arguably the most effective in assessing preconditioning. Macartney et al (40) found that both treatment groups (vaccination and preconditioning) had lower incidence of morbidity compared with calves obtained through conventional auction. Indeed, the benefit attributable to preconditioning was quite large: conventionally sold calves were 4.5 times more likely to be treated than were preconditioned calves. However, the fact that those who made treatment decisions (owners) were not blinded to the origin of the cattle introduces the potential for bias. This concern is arguably bolstered by the absence of difference in mortality between groups. It is conceivable that more aggressive treatment was necessary to prevent mortality in the conventional calves, but it is perhaps more likely that this group of calves was over-treated due to the owners’ expectations of risk. In the Roeber et al (20) study, pen riders were blinded to vaccination status, reducing bias. Results indicated that conventionally sold calves experienced higher morbidity than either preconditioned calves or those vaccinated but not weaned; mortality was also notably higher in the conventionally sold calves, but statistical analysis was not provided (20).

A recent study provided further evidence of the benefits of preconditioning, and also helped define what components are most important for success. Specifically, weaning 45 d prior to shipment (without vaccination) had similar benefits to 45-day weaning with vaccination (43). Both treatments resulted in lower morbidity compared with shipping calves directly to the feedlot or purchasing calves from a salebarn. Since calves from both treatment groups were sent directly to the feedlot, it remains possible that vaccination at or before weaning would benefit calves marketed through more typical avenues, including special sales at sale barns. Nonetheless, the results of this study and that by Roeber et al (20) strongly suggest that weaning and/or vaccinating calves 3 wk or more prior to shipment is beneficial.

Nutritional factors

Receiving diet

Because the feedstuffs consumed in a feedlot are often quite different from those consumed by calves prior to arrival, numerous studies have examined the effect that nutritional factors may have on BRD. Martin et al (4) found that calves fed corn silage had higher morbidity than those fed hay; however, this negative result could be mitigated somewhat by adding grain to the diet. In contrast, Wilson et al (45) found that high grain levels were associated with increased BRD; however, this assessment may have been confounded by other variables, including the source of the cattle and time between arrival and processing. In a more controlled study, concentrates were fed at relatively high levels (72% of ration) without impacting BRD incidence, whereas higher levels (90%) increased morbidity (46). Addition of free choice hay provided no benefit. Fluharty and Loerch (26) found no significant differences in morbidity associated with variation of feedstuffs or varying levels of concentrate (up to 85%). Their findings, however, indicated that the number of treatments required for sick calves increased as the percentage of concentrate increased. Other studies demonstrated a trend for increasing morbidity with increasing levels of concentrate, typically becoming apparent at varying points above 50% of the diet (47). Berry et al (23) maintained a narrow range of concentrate: roughage ratio to assess the impact of varying levels of energy and starch with minimal confounding. This study failed to identify a significant association between energy and starch levels within the ration and BRD. A comprehensive review concluded that diets with increased energy density (concentrates) can be employed for improved ADG without altering incidence of BRD (48). Crude protein (CP) content is another area frequently investigated to determine ideal levels to include in receiving rations. Fluharty and Loerch (26) found no association between CP levels and BRD. In other studies, however, CP appeared similar to concentrate percentage in that increased morbidity occurred as levels increased, with the preferred inclusion level being approximately 14% (47).

Various other nutritional factors have been proposed to reduce BRD, including supplementing potassium, thiamine and other B-vitamins, and rumen bypass protein (3,49,50). None of these consistently reduced disease (3,51). Inclusion of vitamin E in the ration at levels exceeding the recommendations of the National Research Council also has no impact on BRD incidence (52). Researchers have studied the relationship between BRD and various mineral levels as well as the form in which the nutrients are supplied. In 1 study, reduced treatment rates were noted in calves supplemented with metal-complex minerals compared with calves offered no mineral supplement or those receiving inorganic sources (sulfate-complexed) (53). Another study, however, found no difference when comparing organic complexes and sulfate-bound minerals in non-stressed calves (54). Disparity in results of such studies may be due to differences in the mineral status of the calf at time of enrollment; however, this assertion was not supported by a study which found no correlation between mineral deficiency at arrival and treatment for BRD (22). Other studies have examined copper, zinc, vitamin E and selenium, chromium and various combinations of these, evaluating immune function and/or calf performance (for example, ADG, feed/gain ratio). Reviews of the interaction between nutrition and immunity concluded that the inconsistencies in trial results preclude any definitive link between BRD and most specific nutritional factors (47), and “after decades of research, our ability to modify the incidence of BRD through nutritional manipulations seems limited” (48).

Increasing intake

Finally, in considering nutrition’s role in BRD, achieving adequate feed intake may be more important than what is included in the ration. Newly weaned calves may not be accustomed to prepared feeds, eating from feed bunks, and drinking from water troughs; therefore, the use of trainer animals has been suggested to acclimate calves to these activities. Trainers are older cattle, acclimated to the feedlot environment and familiar with the facilities for eating and drinking. Use of cull cows as trainers improved calf health in several, but not all, trials of 1 study. The same investigators found that use of trainer steers was not beneficial, and led to increased disease in the trainer steers (55). Similarly, Gibb et al (18) found that the presence of trainer cows had no effects on morbidity for newly received calves and reduced ADG for calves in the early feeding period.

Prophylaxis/Metaphylaxis

Parenteral antimicrobials

Involvement of bacterial pathogens in BRD suggests mass treatment of at-risk populations could reduce morbidity and mortality. Such mass treatment can be accomplished in 1 of 3 ways-parenterally, in feed, or in water. Administration of parenteral products to calves that are at high risk for BRD (metaphylaxis) has consistently been found to reduce morbidity (19,40,56–60). Products found to be effective include ceftiofur crystalline free acid, florfenicol, tilmicosin and tulathromycin, with conflicting results for oxytetracycline (58,61). Most studies administered these products at feedlot arrival, but pre-shipment treatment was also effective (19,62). Despite the benefit of metaphylaxis, cost and labeling restrictions preclude mass treatment of calves not considered to be at high risk for BRD (63). Calves that have an increased pre-transit body temperature are at greater risk of disease (64); therefore, febrile calves may be considered prime candidates for targeted treatment without administering antimicrobials to the entire group. However, Lofgreen (65) concluded that a treatment system based upon arrival temperatures was not reliable. In another study, selectively treating febrile calves reduced subsequent morbidity compared with no prophylactic treatment, but was less effective than metaphylactic treatment of the entire group (66).

Oral antimicrobials

The benefits of mass administration of oral antimicrobials are less certain. Antimicrobials delivered in water were associated with increased mortality (no distinction was made between which antimicrobials were used) (4). The relevance of this is debatable, however, as results may have been confounded by several factors. First, cattle that had greater early morbidity were more likely to be given antimicrobials in the water. Such groups had higher overall mortality, which may have been due to factors unrelated to administration of the antimicrobials. Second, mortality may have been associated with decreased water intake and subsequent dehydration. This may have occurred due to the number of water sources being reduced so that only those sources with antimicrobials were available, or due to decreased palatability of the water. Alternatively, in the absence of blinding to treatment, some producers may be less diligent in identifying sick calves when antimicrobials are used in the water, because the owner might have assumed that providing treatment in the water decreased the need to treat parenterally (67). Interestingly, the study that found administration in water increased mortality found that antimicrobials in feed reduced morbidity (4). This is in agreement with a study that suggested chlortetracycline and sulfamethazine in feed reduced morbidity at weaning; the results appear convincing but statistical analysis was not performed (68). The inadequacy of these and other trials assessing oral antimicrobials was noted in a meta-analysis (56), which found no well-designed, executed and analyzed studies examining use of mass medication of oral antimicrobials. A study reported after the meta-analysis found that chlortetracycline and sulfamethazine in the ration significantly reduced treatment and mortality in feedlot calves (69). Another recent study reported that chlortetracycline in the feed was as effective as parenteral tilmicosin in preventing BRD (28). That finding was disputed by 2 different studies, where parenteral administration of tilmicosin was preferable to inclusion of chlortetracycline in feed for reducing morbidity (19,70). One of these studies (70) did not include a control group to determine if chlortetracycline provided any benefit, but Frank et al (19) found that calves receiving chlortetracycline in the feed from days 5 to 9 had a similar occurrence of BRD compared with calves not receiving chlortetracycline.

Study considerations

Confounders

Despite the myriad of potential factors studied and associations found, none can provide more than a minor accounting for BRD incidence. Regardless of the design of the study or factors explored, it is common to find that the largest associations are related to farm of origin, destination operation, or both (22,33,40,71). These ill-defined associations not only conceal specific risk factors, they can distort interpretation of those that are identified. Examples of this were alluded to earlier, when discussing pregnancy checking. Certain operations may consistently send pregnant heifers to a feedlot. These operations may also engage in unidentified management practices that increase BRD morbidity. This may result in an apparent association between pregnancy (or pregnancy evaluation) and BRD when 1 does not truly exist. Methods exist to account for potential clustering of disease due to such herd effects, but they have not historically been employed (72). Even when adjustments are attempted, they can be difficult to apply and interpret. Nevertheless, when the study population is composed of animals from different sources or groupings, the potential “herd” effect should be considered and perhaps controlled through a variety of techniques (73). One way to avoid this issue is to consider the group as the unit of interest, rather than individual calves, since most management measures are implemented at the group, not individual level (74). Such a technique would be most meaningful in comparing cohorts that were maintained from birth through feedlot, but could potentially be used for shorter periods of time. Unfortunately, biologically plausible and perhaps real associations that are found at the animal level may not be found at the herd level. For example, Pasteurella species were isolated more commonly from nasal secretions of sick calves than from healthy calves (75), but morbidity was not higher in herds with greater frequency of Pasteurella isolates (76). Additionally, caution must be exercised in drawing conclusions from ecological or group level studies. For instance, what recommendation can be made if there is higher morbidity in a group of calves that included bulls that were castrated after arrival at the feedlot, but the castrated calves were not the ones that were sick?

Nature of bovine respiratory disease: Contagious, clustered, or random?

Another challenge frequently encountered in researching BRD is difficulty in replicating results due to large variability in outcomes. This can occur even when the same people execute the same processes using cattle from the same operation(s) in different years (39,77,78). This variability begs the question as to whether BRD is contagious, if it clusters due to risk factors, or if it occurs randomly. A small-scale study found that risk of disease is not increased by housing calves in a pen with sick cattle (79). This may be interpreted as BRD not being contagious in the classic sense. However, another study found a spike of BRD in bulls already acclimated at a test station when new bulls were introduced (80). Frank et al (57) used a leukotoxin deficient, live attenuated M. haemolytica strain as an intranasal vaccine. This unique organism was subsequently cultured from non-vaccinated calves, suggesting these calves contracted it from vaccinated cattle, albeit at a low and inconsistent rate. This was not the case in a similar study, where no transmission to non-vaccinated calves occurred (81). Further clouding the issue of transmissibility was the finding that antimicrobial resistant strains of bacteria emerged in calves treated for BRD, but there was no evidence of these strains being transmitted to other calves in the feedlot (82). A large, multi-year study concluded that temporal and spatial clustering effects were small (78). This would suggest that, not only is BRD not contagious, but that risk factors would be hard to identify or may not be important, at least in that study population. Another study found cases of respiratory disease often clustered according to truck (a proxy for source) and less frequently by pen (83). This would suggest that BRD is not a random event, and thus is suitable for epidemiologic study. Unfortunately, the investigators were unable to conclude whether contagious or non-contagious factors were more important in disease occurrence.

Molecular techniques have been developed to distinguish among both M. haemolytica (83,84) and P. multocida isolates (85–88). These would appear to be perfect tools in assessing whether there is significant lateral transmission, but no studies have been reported to date. DeRosa et al (89) used ribotyping, serotyping, and antibiograms on isolates from nasal and tracheal swabs. The limited diversity found was suggestive of lateral transmission, although the study was not designed to explore this possibility. Their findings also suggested nasal swab culture can be predictive of the pathogen within the lung (89). This idea was supported by work reported by Godinho et al (90). Using postmortem lung lavage as the gold standard, nasal swabs were found to have 100% positive predictive value for presence of M. haemolytica and M. bovis in the lung (there were too few P. multocida isolates for analysis). Further, random amplified polymorphic DNA polymerase chain reaction (RAPD-PCR) found excellent correlation between lung and nasal isolates, suggesting that culture of nasal passages of sick calves may provide clues as to what strain is present in the lungs. A different conclusion was reached by Allen et al (91), who found only moderate correlation between organisms isolated from nasopharyngeal swabs and those obtained by bronchoalveolar lavage. This conclusion was made in regards to all pathogens examined; nonetheless, good correlation was present between nasopharyngeal and lung isolation when analysis was restricted to P. multocida. Further investigation is warranted to determine if application of molecular techniques to isolates obtained from nasal swabs could provide insight into transmission of BRD pathogens.

Diagnosis of bovine respiratory disease

Perhaps the greatest hindrance to better elucidating factors associated with BRD is difficulty in defining and/or identifying the disease. Postmortem diagnosis, in the absence of prior treatment, is the most definitive means. It is also an objective parameter that could be readily compared across multiple studies, years, or feedlots. Withholding treatment, however, is unacceptable on animal welfare and economic grounds. Postmortem examination following treatment failure has merits, but due to selective pressures exerted by the immune system and antimicrobials, findings in such cases may not reflect etiologies involved in initiating the disease process. Necropsy is also insensitive in detecting those affected, particularly in outbreaks with low case fatality. Thus, when mortality is used as the outcome of interest most studies have low power to detect treatment differences. Mortality rates can also be biased by management decisions of how to manage chronically affected cattle, which may be salvaged, euthanized, or retained but provided no further treatment. Subsequently, morbidity is more commonly used as a measure of BRD incidence than is mortality.

The case definition for BRD varies greatly in the literature. Criteria used for classifying calves as suffering from BRD include elevated body temperature (19,23,64), respiratory signs (nasal discharge, cough, dyspnea, or tachypnea) (80,92), decreased appetite, depression, or some combination of these and other secondary signs (18,91,93). Some investigators have used a clinical scoring technique that incorporates several parameters into a semi-objective score that guides treatment decisions and increases consistency within a trial (21,43,59,91). However, even if the same criteria are used in multiple studies, their application may vary or may be used in different combinations. For example, reported temperature cutoff points range from 103°F (25,26) to 104.5°F (24) or 105°F (69). Many researchers simply accept the judgment of feedlot personnel (13,21,51). Thus a calf is diagnosed with BRD if it is pulled for treatment and found to have no signs referable to a system other than the respiratory tract (22,58).

The accuracy of using treatment as a proxy for morbidity is highly questionable. No correlation was found between treatment and respiratory mortality (94,95), which is an insensitive but specific indicator of BRD. Other researchers have found good to limited correlation between treatment and serological changes (96,97), poor correlation of treatment with bronchoalveolar lavage cytology (96,98), and no or poor correlation between treatment and lung lesions found at slaughter (99,100). In fact, lung lesions are often found in nearly as many calves that were not treated as in those deemed ill (99–102). Thompson et al (101) found calves treated for BRD were 1.5 times more likely to have lung lesions than those not treated, but 69% of those with lesions were never treated. A high prevalence of lung lesions among treated and non-treated cattle at slaughter was also noted by Buhman et al (21); however, these researchers further classified lesions as either severe or minor. Although there was poor diagnostic agreement between previous treatment and presence of any lesion (minor or severe), treated calves were more likely to have severe lesions than non-treated calves (21). Thus, treatment may detect only the most severely diseased cattle.

Several studies found that treated calves have decreased ADG compared to non-treated calves (35,101–103). This is intuitive and suggests that those treated were indeed ill. But other studies have not corroborated this association (20,94,99). Lung lesions at slaughter are consistently correlated with decreased ADG (95,99–101). Two studies found decreased ADG to be correlated with both treatment and lung lesions, with lesions showing a stronger association (100,101). These findings have led some to suggest that ADG would be a better barometer of respiratory disease than field diagnosis (74,94). Because ADG can be impacted by things other than BRD, however, using ADG as the primary indicator of respiratory disease would open studies to many new potential confounders.

Other methods of diagnosis of BRD are actively being researched, including ancillary testing (104), metabolic profiling (43,64,105), and improved field diagnostics (106). Attempts to identify a single indicator of disease have been disappointing. Serum cortisol levels have been measured in BRD studies, but cortisol concentration does not distinguish between disease and stress, and is difficult to interpret due to its diurnal variation and rapid changes in circulating levels (105). Similarly, acute phase proteins have generally been unreliable for identifying cattle suffering from BRD, although haptoglobin has shown promise (107,108). Multi-modal approaches appear to offer potential for improvement in diagnosis. For example, a panel of analytes, including acute phase proteins, metabolites, and microminerals was able to differentiate between stress, infection, and stress combined with infection (105). It was not possible, however, to establish a baseline value for any of the analytes that could distinguish between “normal” and “diseased” animals. It remains to be seen whether such approaches are viable for use in epidemiologic studies.

Conclusion

It is evident that BRD is, and will continue to be, an extremely costly condition. Although much research has been done regarding its prevention, there are only a few conclusive findings. Preconditioning offers some benefit (at least to the purchaser), but efficacy is variable. Weaning prior to sale is perhaps the most important component of preconditioning, although vaccination prior to sale may offer benefits. Vaccination after arrival appears to have limited value. The practice with the clearest benefit is metaphylaxis, and yet the costs, both monetarily and in terms of potential antimicrobial overuse, preclude its routine practice for all cattle. Identifying reasons for tremendous variability in research results may improve our ability to more accurately identify the contributions various factors make to disease occurrence as well as effectiveness of mitigation strategies.

It is imperative that researchers and practitioners recognize limitations of the current knowledge, even while attempting to do all possible to reduce the impact of BRD on the beef industry. It would appear that the goal warranting greatest attention is improved diagnosis. Until an objective, consistent criterion can be established for BRD, it is unlikely that universal findings will be identified. A second question worthy of consideration is heterogeneity within the agents associated with BRD. Given the variability in morbidity and mortality that cannot be readily ascribed to host or environmental factors, it seems reasonable to investigate whether some strains within a given bacterial species are more capable of causing disease than others. Such research may also provide insight into virulence mechanisms, improving prevention and treatment options. CVJ