Assessment of within-herd seroprevalence of Anaplasma marginale antibodies and associated decreased milk production in an Iowa dairy herd

Background

In late 2010 a dairy located in northern Iowa was diagnosed as being infected with bovine anaplasmosis. Before this diagnosis, herd mortality data indicate annual death loss among all cows at the site increased from 124 (16.7%) and 121 (16.0%) in 2006 and 2007, respectively, to 182 (27.5%) cows in 2008. Death loss peaked in 2009, with 229 deaths (32.9%), and then declined. A 2014 National Animal Health Monitoring System survey (USDA, NAHMS, 2014) reported that average annual cow mortality was 5.6% across US dairy herds. This period of elevated death loss coincided with decreased milk production from January 2008 to January 2009. The first diagnosis of anaplasmosis was made in October 2010. The dairy herd was open, implying that animals were purchased and introduced without prepurchase testing or quarantine. Records documenting the origin of new animal purchases were not kept, but a herd vaccination program was in place. Both calves and adults received standard cattle immunizations, which included vaccines for bovine viral diarrhea virus (BVDV). Records indicate that calves received vaccinations against BVDV infection at 1 wk of age, 6 to 8 wk of age, and at 10 to 12 mo of age. Dry cows were vaccinated against BVDV at 60 d before calving. Lactating cows were vaccinated against BVDV at between 35 and 41 DIM. All lactating cows received oxytocin injections twice daily at the time of milking. Syringes and needles were reused among animals receiving oxytocin, and visible blood contamination was observed in the syringes and oxytocin bottles during a farm visit.

Mortality data were obtained from records archived in on-site dairy management software (Dairycomp 305, Valley Agricultural Software, Tulare, CA). Milking and DIM records were obtained from DHIA archives. Unlike animal health data, DHIA milk production data are collected once a month by an independent observer. Changes in DHIA milk production data and death loss over time were examined to establish the progression of the outbreak. Veterinary billing records were also examined as part of the herd health review.

Beginning in 2011 anaplasmosis infection status (positive or negative for Am antibodies) was determined using a commercially available competitive ELISA (cELISA; catalog No: 283-2, VMRD, Pullman, WA) designed to provide results that will give guidance about the presence of Anaplasma infection in cattle. This test kit was approved by the USDA, and the test was conducted in an American Association of Veterinary Laboratory Diagnosticians–accredited facility at Iowa State University. As per the manufacturer (VMRD), the test kit has a diagnostic sensitivity of 100% and a specificity of 99.7%, with a cutoff point of 30% inhibition. It has been reported that this test kit can be cross-reactive to Anaplasma phagocytophilum (Dreher et al., 2005) and a genotype of Ehrlichia (Al-Adhami et al., 2011). However, as there has never been a reported case of naturally occurring A. phagocytophilum infection in US cattle (Hairgrove et al., 2015), and only one reported instance of naturally occurring Ehrlichial infection among cattle in Canada (Gajadhar et al., 2010), the VMRD kit was deemed appropriate for the survey.

Serology

Blood samples were obtained from 799 of the Iowa dairy cows throughout 2011 in 24 separate accessions. Blood samples were submitted by 1 of 2 licensed veterinarians, who had professional relationships with the dairy owner. Samples were analyzed for Am antibodies using cELISA, which assumed a cutoff of 30% inhibition for a sample to be considered positive for antibodies. Although serology data became available throughout 2011, animals remained intermingled regardless of serostatus and were managed under the same production conditions. The number of animals tested was not entirely comprehensive as cows were tested at different time points and animals were bought and sold during the sampling process. Documents relating to the case indicate that sampling days occurred approximately every other week, and timing was based on the discretion and availability of a licensed veterinarian familiar with the operation. Sampling occurred, at least in some cases, when cattle were being caught and tested for other diseases (e.g., brucellosis). Each animal contributed a single data point to the analysis.

Statistical Analysis

Microsoft Excel (Excel for Mac Version 16.38, Microsoft Corporation, Redmond, WA) was used for data compilation and descriptive statistics. To examine the association between anaplasmosis serological status and milk production, data were analyzed in JMP (SAS Institute Inc., Cary, NC) using a Mixed Effects statistical model. Animal nested in anaplasmosis serological status (positive or negative) was designated as a random effect, with anaplasmosis status, year, and their interaction designated as Fixed Effects in the model. To account for differences in DIM, milking frequency, season of calving, and cow age, mature equivalent (MEq) lactation data calculated and reported by DHIA was used for comparing milk production between positive and negative cows. Statistical significance was designated a priori as P < 0.05.

Diagnostic Testing

An investigation into the prevalence of Am in the 799 cows sampled from an Iowa dairy was conducted in 2011. The cELISA results from this survey indicated 38% of the animals were positive for anaplasmosis. This number does not take into account the animals dying from the disease before the diagnosis of anaplasmosis in late 2010. These results are consistent with what is typically encountered in herds experiencing an outbreak of anaplasmosis because approximately two-thirds of the herd would be considered susceptible to the disease, whereas one-third of the herd is infected. As infected cattle serve as reservoirs of infection, bacteria can be vectored to local naïve cattle primary through contaminated equipment or arthropods. This endemic instability favors disease outbreaks as occurred at the Iowa dairy.

Needle Transmission

Outbreaks of anaplasmosis can occur due to a combination of different factors. These include a lack of a disease control program, a high ratio of susceptible cattle relative to cows persistently infected in a herd, and the prevalence of vector/fomite transmission (Gill, 1994). The ratio of susceptible to unsusceptible cattle in this case was variable, but testing revealed that the majority of animals in this herd were naïve. Contributing to observed seroprevalence was the lack of a quarantine or testing process for newly purchased animals before being introduced to the herd.

Natural vectors have been implicated in the spread of Am and include members of the genera Tabanus (Ewing, 1981), Psorophora (Ristic, 1968), Stomoxys (Baldacchino et al., 2013), and Dermacentor (Kocan et al., 1981). Species of Tabanus (Sutton and Millspaugh, 1950), Stomoxys (Raun and Casey, 1956), Psorophora (Dunphy et al., 2014), and Dermacentor (Lingren et al., 2005) are known to exist in Iowa and could have contributed to transmission at the dairy. In this case, however, it is apparent that a lack of comprehensive biosecurity measures was likely a significant contributor to Am spread from seropositive to seronegative cows.

Blood-contaminated fomites, such as needles, dehorning saws, nose tongs, tattooing instruments, ear tagging devices, and castration instruments, have also been shown to contribute to disease transmission (Kocan et al., 2003). Needles as a potential vector appear particularly likely as it was discovered that needles were shared among cattle when oxytocin was injected for milk ejection. One study that considered iatrogenic transmission of Am concluded that, when exposed to contaminated needles, naïve steers exhibited disease 60% of the time following a single exposure (Reinbold et al., 2010). As Am positive animals were commingled with those that were Am negative, the reuse of needles in the milking parlor likely served as a point of disease transmission to naïve cows.

Milk Production Analysis

Milk production data were obtained from DHIA records archived in on-site dairy management software (Dairycomp 305, Valley Agricultural Software). Review revealed that before January 2008, herd milk production peaked at 665,863 kg in May 2007, with a nadir of 529,760 kg in October 2006. After January 1, 2008, milk production peaked approximately 90,718 kg lower than in 2007, with trough production reduced by approximately 136,078 kg to 403,194 kg in February 2010. Rolling herd average milk production decreased by 880 kg/cow from 2008 to 2010. This represented an 8% decrease in milk production.

A statistical analysis of the MEq milk production per cow by anaplasmosis status over 3 yr revealed evidence of serological status (P = 0.005) and a year-by-serological status interaction (P = 0.0297). Specifically, cows that tested positive for anaplasmosis on the cELISA test in 2011 produced on average 1,389 kg less per year than cows that tested negative. This difference over time from 2011 to 2013 is presented in Figure 1. The difference in milk production between positive and negative cows was not significant in 2011 (P = 0.427) but was highly significant in 2012 (P = 0.0041) and 2013 (P = 0.0351). In 2012 Am positive cows produced, on average, 10.4% less milk on a MEq basis than those that were Am negative. In 2013 positive cows averaged 15.3% less milk than negative animals. Ordered differences between positive and negative cows are shown in Table 1.

1. Download : Download full-size image

Figure 1. The DHIA recorded mature equivalent (MEq) milk production by anaplasmosis infection status. Error bars indicate SE.

Table 1. Ordered differences in milk production between anaplasmosis positive and negative status

Ordered difference			Difference (kg)	SED1	Lower confidence limit	Upper confidence limit	P-value
Level 1		Level 2
Negative, 2013		Positive, 2013	2,174.85	1,021.71	154.14	4,195.56	0.0351
Negative, 2012		Positive, 2012	1,677.35	579.07	537.53	2,817.17	0.0041
Negative, 2011		Positive, 2011	315.87	397.02	−465.88	1,097.62	0.427

1

SED = SE of the difference.

Though there is a paucity of literature on the subject of milk production during an anaplasmosis outbreak, a few reports have supported these observations. Pazinato et al. (2016) reported that cows identified as seropositive for Am produced, on average, 3 L of milk less per day than identically managed cows found to be seronegative. Research conducted on a dairy in an area considered endemic for bovine anaplasmosis showed that low-producing dairy cattle (i.e., <1,500 kg of milk per annum) were 3.9 times more likely to be seropositive for Am than high-producing animals (i.e., >3,000 kg milk per annum). In the same study, primiparous dairy cattle were found to have an 88% greater chance of being seropositive than multiparous herd mates (da Silva and da Fonseca, 2013). An unspecified decrease in milk production was reported in Turkish dairy cows subsequently tested and found to be seropositive for Am (Birdane et al., 2006; Aktas and Ozubek, 2017). McDowell et al. (1964) reported an average loss in lactation yield of 26% in milk and 31% in milkfat during the clinical phase of disease. These results are consistent with our findings of a decrease of 10.4 to 15.3% when comparing seronegative to seropositive cows. However, our survey focused on the long-term effects of positive serological status as opposed to the effects of clinical disease.

One potential limitation to the survey is the unknown effect of bovine leukosis virus (BLV) on milk production in this case. Evidence suggested that at least some animals at the Iowa dairy had developed lymphoma as a result of BLV infection. Previous studies have shown that that BLV-positive cows produce from 3% (Ott et al., 2003) to 11% (D’Angelino et al., 1998) less milk than did BLV-negative animals. However, it is noteworthy that cows in these reports were not screened for Am. It is possible that the decrease in production resulted from a combination of factors (i.e., animals could have been positive for both Am and BLV). As the BLV status of animals subject to this survey was unknown, this relationship could not be considered.

Cattle were routinely vaccinated as part of the herd management program on the dairy, and animals did not exhibit signs of infection with BVDV during the period described. However, it is possible that latent BVDV could predispose cattle to developing clinical anaplasmosis. A report on Hungarian dairy cattle demonstrated that if susceptible animals are infected with both Am and BVDV at roughly the same time, the immunosuppressive effect of BVDV will support the progression of Am infection to the point of clinical disease (Szabara et al., 2016). As animals were not routinely screened for BVDV, the potential for simultaneous infection cannot be ruled out completely.