|Home | About | Journals | Submit | Contact Us | Français|
A prospective cohort observational study was conducted to investigate the seroprevalence of Anaplasma marginale in Iowa feedlots and its association with morbidity, mortality, and treatment costs. Blood samples were taken from 659 calves from 31 consigners at processing and classified as seropositive to A. marginale using a competitive enzyme-linked immunosorbent assay (cELISA) with a 30% cutoff. Health and production parameters were modeled by A. marginale serostatus with mixed model regression analysis. The apparent prevalence of seropositive cattle was 15.17% (100/659). When the cELISA positive cutoff was at 42% inhibition, the apparent prevalence was 5.00% (33/659). There was no significant association between A. marginale serostatus and production parameters; however, seropositive status had a weak positive association with undifferentiated fever (P = 0.17). Although prevalence of anaplasmosis in Iowa feedlots is higher than reported in Montana-sourced calves arriving in Canadian feedlots, this was not associated with increased production costs.
Séroprévalence d’Anaplasma marginale dans deux parcs d’engraissement de l’Iowa et son association avec la morbidité, la mortalité, les paramètres de production et les caractéristiques des carcasses. Une étude prospective par observation de cohortes a été réalisée pour faire enquête sur la séroprévalence d’Anaplasma marginale dans des parcs d’engraissement de l’Iowa et son association avec la morbidité, la mortalité et les coûts de traitement. Des échantillons de sang ont été prélevés de 659 veaux auprès de 31 consignataires et classés comme séropositifs pour A. marginale à l’aide d’un dosage immuno-enzymatique par compétition (cELISA) avec un seuil de coupure de 30 %. Des paramètres de santé et de production été ont modélisés en fonction du statut sérologique envers A. marginale avec une analyse de régression d’un modèle mixte. La prévalence apparente du bétail séropositif était de 15,17 % (100/659). Lorsque le seuil de coupure cELISA se situait à une inhibition de 42 %, la prévalence apparente était de 5,00 % (33/659). Il n’y avait aucune association significative entre le statut sérologique pour A. marginale et les paramètres de production; cependant, le statut séropositif présentait une association positive faible avec une fièvre indifférenciée (P = 0,17). Même si la prévalence d’anaplasmose dans les parcs d’engraissement de l’Iowa est supérieure à celle signalée chez les veaux provenant du Montana arrivant dans les parcs d’engraissement canadiens, cela n’était pas associé à des coûts de production accrus.
(Traduit par Isabelle Vallières)
Anaplasmosis, caused by the rickettsial hemoparasite Anaplasma marginale, is one of the most prevalent tick-transmitted diseases of cattle worldwide (1–3). Acute anaplasmosis is associated with mortality and sporadic abortions in beef and dairy cattle and is characterized by fever, anemia, and icterus following a 3- to 6-week incubation period. Economic loss to the United States (US) livestock industry as a result of anaplasmosis is estimated to be > $300 million/y (3). The Canadian Food Inspection Agency (CFIA) has concluded that anaplasmosis is currently not found in Canada (4). Direct production losses associated with the establishment of anaplasmosis in Canada, however, have been estimated to be $12 to 36 million in 2003, with a CFIA cost to prevent the spread of anaplasmosis of $3 million (4).
Cattle that recover from acute anaplasmosis, including those treated with recommended doses of tetracyclines, develop persistent infections characterized by cyclic rickettsemia that ranges from 102 to 107 infected erythrocytes/mL of blood and occurs at intervals of approximately 5 wk (5–7). Although infected erythrocytes are not always detectable in stained blood films during these cycles, persistently infected cattle serve as reservoirs of infection for mechanical and tick-borne transmission (8–10). Processing cattle on arrival at feedlots is therefore a potential risk factor for iatrogenic transmission of anaplasmosis via contaminated injection needles, implant devices, and other animal husbandry equipment. Currently serological or molecular diagnostic tests are the only reliable means to identify cattle persistently infected with A. marginale (11,12).
A commercially available competitive enzyme-linked immunosorbent assay (cELISA) is currently the serological test preferred by regulatory authorities for identifying anaplasmosis carriers (12). This test is reported to have a sensitivity of 96% and specificity of 98% when used to identify persistently infected cattle (13). Results are measured relative to a standardized positive control sample with 30% inhibition representing the seropositive cutoff in the US and 42% representing the seropositive cutoff in Canada (13,14). A previous study of Montana-sourced weaned calves and yearling cattle entering Alberta feedlots found that the apparent prevalence of antibodies to A. marginale determined using the cELISA test was < 2% (14). However, clinical anaplasmosis is considered uncommon in the northern States such as Montana, Idaho, Washington, and North Dakota (4).
The prevalence of anaplasmosis in feedlots in A. marginale endemic regions of the US and the association between anaplasmosis serostatus and production parameters has not been reported. This information is important because vaccination and other invasive management procedures at feedlots may be risk factors for iatrogenic transmission of anaplasmosis especially where carrier calves are co-mingled with naïve animals. The purpose of this study was to investigate the seroprevalence of A. marginale in cattle arriving at 2 Iowa feedlots and to evaluate associations between serostatus and production parameters, morbidity, mortality, and treatment costs.
This prospective cohort observational study was conducted in 2 western Iowa commercial feedlots in Fall 2002. Animals (n = 659) from 31 consigners were enrolled over 5 separate days. The cattle were spring and summer born calves from Iowa. According to the feedlot protocol, all calves entering the feedlot had been previously vaccinated at least once against BVD, IBR, PI3, BRSV, Histophilus somni, Pasteurella multocida, Mannhemia haemolytica, and Clostridium sp. Calves received a modified live IBR, BVD, BRSV, and PI3 vaccine, growth promoting implant, and anthelmentic pour-on at processing which occurred on the day of arrival at which time blood samples were collected in 10 mL serum separation tubes. All samples were transported to the laboratory, processed, and stored at −20°C within 24 h of collection, except for 1 group of 100 samples which were refrigerated for 5 d prior to freezing.
A commercial cELISA (Anaplasma Antibody Test Kit, cELISA, VMRD, Pullman, Washington, USA) which is based on recombinant Anaplasma MSP-5 antigen produced by plasmid- transformed Escherichia coli, was used in accordance with the method described by the Office international des épizooties (OIE) and recommended by the manufacturer (12,13). Samples that had inhibition of < 30% were recorded as negative results, whereas samples that had inhibition of ≥ 30% were recorded as positive. Serum was also tested for antibodies to Neospora caninum (HerdChek Anti-Neospora caninum Antibody Test Kit; IDEXX Laboratories, Westbrooke, Maine, USA) and bovine viral diarrhea virus (BVDV) (BVD Virus Antigen Detection ELISA; Syracuse Bioanalytical, East Syracuse, New York, USA) based on previous reports suggesting an association between antibody status and production losses (15,16). For N. caninum, a cut-off of 0.40 S/P for a positive test result was used; test performance parameters include a sensitivity range of 97.56% to 100% and specificity range of 93.75% to 98.53%. For BVDV, a cut-off of 0.30 S/P for a positive test result was used; test performance parameters include a sensitivity range of 93.6% to 100% and specificity of 100%.
Morbidity and mortality was recorded by the feedlot staff using a record system designed for accounting of treatment and feed costs. A treatment was considered a disease event if a calf was treated with antibiotics, was lame, showed neurological signs, and/or had a temperature greater than 40°C (104°F). Separate events were recorded if 4 d had passed between treatments. Disease categories in which at least 1 event was recorded by feedlot staff included: bovine respiratory disease, undifferentiated fever, digestive disease, and other. Performance data were recorded by the commercial feedlots. Carcass data was determined using USDA grades and was provided to consigners by abattoir facilities for 397 study animals after about 85 d on feed. Performance and carcass variables are listed in Table 1.
For all analyses the unit of concern was the calf. All serological data were recorded as seropositive or seronegative based on the mean of the duplicate serological tests > 30% inhibition. The apparent anaplasmosis prevalence was calculated as the number of animals where the mean cELISA results were positive (≥ 30% inhibition) divided by the total number of animals tested. The manufacturer reported a test sensitivity of 96% and specificity of 98% when a 30% cutoff for positive samples is used. The apparent anaplasmosis prevalence based on the cELISA was adjusted for the reported test sensitivity and specificity, as previously described (17), using the following equation:
Agreement between the cELISA results obtained following duplicate testing for each sample (test 1 and test 2) was assessed by calculating a Kappa statistic (κ) (18,19). Data were converted to a binary format (0 = no disease and 1 = disease) by use of a cutoff value of 30% inhibition, as described previously. Thereafter, data were compared by contingency analysis by use of a 2 × 2 table in which κ represented the actual agreement between the tests divided by the potential agreement beyond chance.
Percent seropositive and seronegative animals with 95% exact confidence intervals (95% CI) were calculated using commercial statistical software (SAS 8.1, SAS Institute, Cary, North Carolina, USA). Morbidity related variables included: “treated,” a dichotomous variable representing at least 1 treatment period during the feedlot period with yes and no responses, “number of times treated” a count variable with values 0–9, and “disease category” a polychotomous categorical variable. The recorded mortality variable was “dead,” which represented animals that died during the feedlot period.
A t-test was used to describe the univariate association between Anaplasma serology status and the continuous variables listed in Table 1, that included average daily gain, feed to gain ratio, delivery weight, and hot carcass weight. A chi-squared test for proportion was used to compare the association between Anaplasma status and the categorical variables sex, mortality, morbidity, and BVDV status.
Based on the univariate analysis, variables were selected for inclusion in all initial models if Wald statistic P-values for continuous variables and Fisher’s exact test two-tailed P-values were < 0.25. Several variables, including sex and delivery weight, were forced into the initial models because of their potential confounder status.
To test the null hypothesis that A. marginale seropositive Iowa feedlot cattle had similar morbidity and mortality risk compared to seronegative cattle, 2 generalized linear models with a logit link (mixed model for discrete data) (PROC MIXED, %GLIMMIX macro, SAS 8.1, SAS Institute) were created. The outcome variables for the 2 models were “dead in feedlot period” or “treated during feedlot period.” For each model, the initial explanatory variables were A. marginale serostatus, BVDV PI status, gender and delivery weight, as fixed effects: consigner group was a random effect. For the mortality model “treated” was included as a fixed effect. Backward elimination was then used to eliminate variables from the initial model using the exit criteria of Wald statistic with P < 0.05.
To test the null hypothesis that the occurrence of chronic disease was not associated with A. marginale serological status a generalized linear model with a logit link was also used. The outcome of interest was the occurrence of chronic disease. For this hypothesis 4 definitions of chronic disease were evaluated. The definition of chronic disease became more restrictive with each model (see Table 2). In the 1st model animals treated more than once were defined as chronic. In the 2nd model, animals treated more than 2 times were defined as chronics, etc., up to our most restrictive definition of chronic disease in which animals treated more than 4 times were considered chronic. For all models, the explanatory variable of interest was A. marginale serostatus, which was treated as a fixed effect. Other variables included in the models were N. caninum serostatus, BVDV PI status, gender, and delivery weight as fixed effects and consigner group as a random effect.
To test the null hypothesis that A. marginale seropositive Iowa feedlot cattle had similar treatment costs, production parameters, and carcass traits compared to seronegative cohorts, analysis consisted of a generalized linear model for continuous variables (PROC MIXED, SAS 8.1, SAS Institute). Treatment costs were determined by feedlot managers and included the cost of drugs, supplies, and labor as charged to consigners. As with morbidity and mortality models, consigner group was included as a random effect and A. marginale serostatus, N. caninum serostatus, BVDV PI status, gender, delivery weight, and morbidity were included as fixed effects. Any variable forced into the model Univariate analysis and backwards elimination were used as described previously. Potential confounding variables, consignor group and gender were included in final models.
The apparent prevalence of A. marginale seropositive cattle based on a mean cELISA ≥ 30% inhibition was 15.17% (100/659) (95% CI: 12.52% to 18.14%). When prevalence was adjusted for test sensitivity (96%) and specificity (98%) at a 30% cutoff the adjusted prevalence was 14.85%. Seventy-four percent (23/31) of consigner groups had at least 1 seropositive animal. When the CFIA cutoff of 42% inhibition is applied to our results, the apparent prevalence is reduced from 15.17% to 5.00% (33/659) with a decrease in adjusted prevalence from 14.85% to 4.88%.
Agreement between the duplicate cELISA tests based on the κ statistic was 0.74 ± 0.036 with 7 calves testing positive on test 1 and negative on test 2 and 39 animals testing negative on test 1 and positive on test 2. Within consigner groups, prevalence of A. marginale seropositive cattle ranged from 0% to 60%. Two animals were identified as persistently infected with BVD, neither of which was seropositive for A. marginale antibodies. When the cELISA test cutoff was increased from 30% to 42% inhibition, the κ statistic increased to 0.77 ± 0.053. However, when test results interpreted using a 30% inhibition cutoff were compared with results interpreted with a 42% inhibition cutoff, the κ statistic was 0.46 ± 0.052.
Overall treatment rates, morbidity, and mortality for A. marginale seropositive and seronegative animals are shown in Table 2. The overall morbidity of the 664 calves sampled was 30.05 % (95% CI: 26.57% to 33.71%) and the mortality was 2.43% (95% CI: 1.39% to 3.91%). Morbidity among A. marginale seropositive calves was 33% (95% CI: 23.92% to 43.12%) while the morbidity among seronegative calves was 29.52% (95% CI: 25.76% to 33.49%) (P = 0.484). With potential confounding variables included in the regression models, no significant associations between A. marginale serostatus and morbidity or mortality variables were found.
The distribution of chronic treatment between A. marginale positive and negative animals is shown in Table 3. There was a total of 83 calves that received 2 or more treatments, representing 12.5% of the total number of calves sampled. Only 9 (11%) of these calves were seropositive for anaplasmosis while the remaining 74 (89%) were seronegative for anaplasmosis (P = 0.325). Since the distribution of treatments between seropositive and seronegative animals was similar, there was no association between A. marginale serostatus and individual treatment costs in this study (Table 4). Furthermore, there was no significant difference in days on feed, feed to gain, and overall average daily gain between Anaplasma seropositive and seronegative calves.
Anaplasma marginale serostatus had no association with any of the chronic disease variables (P = 0.325 to 1.000) or respiratory disease (n = 151, P = 0.80). However, A. marginale seropositive status had a weak positive association with undifferentiated fever (n = 39, P = 0.17). Based on the univariate and multivariate analyses, production and carcass parameters were not associated with A. marginale serostatus of feedlot animals. Dressing percentage, hot carcass weight, yield grade, and percent retail product were not significantly different between seropositive and seronegative animals. Descriptive statistics and parameter estimates for production and carcass variables are included in Tables 4 and and5,5, respectively.
The results of the present study suggest that the apparent prevalence of anaplasmosis on arrival in Iowa feedlots (15.17%) is considerably higher than the prevalence of 1.93% previously reported for yearling cattle imported into Alberta feedlots from Montana in 2001 (14). A follow-up study in the same region reported prevalence rates of 1.82% and 1.35% in 2002 and 2003 respectively (20). Anaplasmosis has been reported in all 48 contiguous states in the United States although clinical anaplasmosis is uncommon in the northern states such as Montana, Idaho, Washington, and North Dakota (14). Apparent seroprevalence reports range from very low in northern states (1.93% in Montana) to moderately high in the Midwest (7.6% in Illinois) (21), Gulf (3.8% to 11.2% in Louisiana) (4), and southern states (4.7 to 17% in Oklahoma) (22). There is no published literature on serological surveys of the prevalence of anaplasmosis in feedlots in the southern US. To our knowledge, this is the first study to report the prevalence of anaplasmosis in calves at processing in US feedlots and to examine associations with morbidity, mortality, production parameters, and carcass traits.
The number of calves potentially exposed to anaplasmosis during the feeding period could not be quantified in the present study as sampling only occurred on arrival. This is an important consideration in interpreting the results of the present study given that serological status on arrival was used as the explanatory variable for mortality and morbidity events that happened subsequently. Vector-borne transmission of anaplasmosis is considered uncommon in feedlots and is unlikely to have occurred in winter in Iowa since this is the non-vector season. However, iatrogenic transmission via contaminated equipment at processing is a consideration. Calves may therefore have been misclassified as seronegative impacting subsequent associations between serological status and production parameters.
The κ statistic measures the agreement between 2 tests on a scale from 0 to 1. When applied to diagnostic test results, common interpretations of κ are that < 0.2 is slight agreement, 0.2 to 0.4 is fair agreement, 0.4 to 0.6 is moderate agreement, 0.6 to 0.8 is substantial agreement, and > 0.8 is almost perfect agreement (19). In this study, a κ statistic of 0.74 ± 0.036 (30% cutoff) was calculated indicating substantial, but not perfect, agreement between the 2 duplicate ELISA tests. One explanation for the discrepancy between tests is that chance-corrected measures of association require test results to be converted to discrete dichotomous outcomes (positive or negative) based on a pre-determined cutoff value. This implies that the agreement between tests may change depending on which cutoff value is selected to determine serological status.
In the United States, the cELISA test manufacturer (VMRD) suggests a 30% inhibition cutoff to determine Anaplasma positive status which was therefore used to determine serostatus in the present study (13). However, the Canadian Food Inspection Agency has previously selected a 42% inhibition cutoff in serological surveys of Canadian cattle using the VMRD test (14). When the 42% cutoff was applied to our results, the apparent and adjusted prevalence was reduced from 15% to 5%. This is consistent with a reduction in apparent prevalence from 1.93% to 0.73% noted when cELISA results from Montana-sourced calves were interpreted using a 42% inhibition cutoff (14).
When the cELISA test cutoff was increased from 30% to 42% inhibition, the κ statistic still indicated substantial, but not perfect, agreement between the 2 duplicate ELISA tests. However, when test results interpreted using a 30% inhibition cutoff were compared with results interpreted with a 42% inhibition cutoff, the κ statistic indicated only moderate agreement between tests. This finding suggests that increasing the cutoff has a substantial effect on the interpretation of anaplasmosis cELISA test results. This has implications for disease certification and international trade in livestock, especially if territories apply different cutoff levels for results obtained from the same diagnostic test. Furthermore, this indicates that cELISA results in the 30% to 40% inhibition range may be inconclusive and the tests should likely be repeated. We suggest that diagnostic laboratories report percent inhibitions with the cutoff specific test interpretation to allow clinicians to determine if animals have been accurately classified as anaplasmosis positive or negative.
Procedures such as vaccination, ear tagging, hormone implanting, dehorning, and castration performed soon after cattle arrive at feedlots are collectively known as processing (22). Blood-contaminated needles and surgical instruments are recognized methods of transmission of anaplasmosis. In a review of several reports of anaplasmosis outbreaks attributed to iatrogenic infections, Dikmans referenced a case involving 1500 animals associated with exposure to contaminated dehorning and horn tipping equipment (23). Stiles similarly published a report of 105 cases of anaplasmosis that occurred following vaccination (24). This observation has subsequently been confirmed experimentally (8,25). The apparent prevalence of anaplasmosis in calves arriving at feedlots reported herein emphasizes the importance of using clean equipment to minimize disease transmission at processing.
The term “undifferentiated fever” (UF) is used to describe feedlot cattle which have signs of depression and pyrexia in the absence of abnormal clinical signs referable to a specific organ system other than the respiratory system (26). The reason for the weak association between anaplasmosis serostatus and UF reported in the present study is not known. Since the number of calves enrolled in this study was not sufficient to detect a statistically significant association between these parameters, a larger investigation is required to elucidate this. With the proportions of disease positive and negative animals in this study a minimum sample size of 985 animals would be required for a statistical power of 0.80. Fever is considered the first recorded sign of anaplasmosis and fever in excess of 40°C usually persists throughout the period of increasing parasitemia (27). In acute anaplasmosis, phagocytosis of parasitized erythrocytes causes severe anemia, resulting in increased heart and respiratory rates (27–29). It is therefore possible that some cases of UF may be attributable to clinical anaplasmosis rather than BRD.
Calves with UF attributable to anaplasmosis may therefore be inappropriately treated with antimicrobials that are only effective against BRD pathogens. Macrolides, florfenicol, and tetracyclines are the primary antimicrobials used as part of the initial treatment of calves presenting with UF (30). Of these, only the tetracyclines are effective against acute anaplasmosis (29). It is also noteworthy that anaplasmosis infections are not sterilized at the usual recommended tetracycline doses (31). These cattle will remain persistently infected and may thus serve as lifelong reservoirs of infection in the feedlot (23). However, since A. marginale infection in calves less than 1-year-old is generally self-limiting and rarely fatal (28), the resolution of UF in these cases may in some cases be incorrectly attributed to a response to antimicrobial therapy.
Although anaplasmosis outbreaks in calves are uncommon, there are reports that cattle on a higher plane of nutrition develop more severe anaplasmosis than animals maintained on a lower energy plane (32,33). In 1 study, Thompson et al (34) reported 122 cases of clinical anaplasmosis in 2036 feedlot calves aged between 5 and 10 mo. Regression analysis of the data found that calves diagnosed and treated for anaplasmosis during the late finishing period (d 35 to slaughter) had a reduction in average daily gain (ADG) of 96 g (P < 0.001). In addition to a reduced ADG, prehepatic icterus associated with tick-borne infections is considered a major cause of carcass condemnation in countries where hemoparasitic diseases are prevalent (35). Consequently, although no significant association between anaplasmosis serostatus and performance was found in the present study, clinical anaplasmosis has significant economic implications in feedlot operations.
In conclusion, the results presented here indicate that the apparent prevalence of anaplasmosis in Iowa feedlots is higher than previously reported in Montana-sourced calves arriving in Canadian feedlots. These data emphasize the importance of implementing biosecurity and biocontainment strategies to prevent iatrogenic transmission of anaplasmosis during processing. Calves that were seropositive for A. marginale did not have higher morbidity, mortality, or individual treatment costs than seronegative cohorts. Although a weak association between anaplasmosis serostatus and undifferentiated fever was found, additional studies are needed to characterize this association further.
Dr. Coetzee was supported in part by the Lloyd Endowed Professorship and Fort Dodge Fellowship. CVJ
This study was funded by the Iowa Livestock Health Advisory Council.
Use of this article is limited to a single copy for personal study. Anyone interested in obtaining reprints should contact the CVMA office (gro.vmca-amvc@nothguorbh) for additional copies or permission to use this material elsewhere.