Brucella abortus RB51 Persistent Infection and Milk Shedding in Vaccinated Dairy Cattle: A Public Health Risk

ABSTRACT

Introduction

Brucellosis, caused by a group of bacteria within the genus Brucella, is a zoonotic disease of significant public health concern. The RB51 strain vaccine is used to combat cattle brucellosis but has potential health consequences for exposed humans.

Methods

A retrospective review of NVSL's records prior to 2021 allowed identification of six cases of RB51 strain Brucella in the milk of dairy cattle through culture, followed up by PCR and whole genome sequencing (WGS).

Results

Six properly vaccinated dairy heifer calves actively shed strain RB51 in their milk as adults, years after vaccination. Two of these cows are linked to three human strain RB51 brucellosis cases (one cow to one human, one cow to two humans). All human cases were detected following consumption of unpasteurised (‘raw’) dairy products, highlighting the risk of raw dairy consumption.

Conclusions

These cases demonstrate that milk‐borne RB51 strain Brucella has been responsible for multiple human cases of Brucella. Dairy products from cows vaccinated with RB51 should not be consumed raw. The continued and widespread use of the RB51 vaccine should be evaluated critically, and authorities and veterinarians should revisit regulations and vaccination practices to ensure they are epidemiologically justified, particularly in regions or herds without documented brucellosis risk.

Impacts

• Dairy calves appropriately vaccinated against RB51 may become persistently infected with the vaccine strain and shed the bacteria in their milk as adults, years after vaccination.

• RB51 is a foodborne illness risk in raw milk; unlike many bacterial infections, it is resistant to common antibiotics and difficult to diagnose.

• Authorities and veterinarians should evaluate whether RB51 vaccination is justified in their region and limit its use to populations with a demonstrated epidemiological need rather than applying it as a blanket requirement.

Abbreviations

RB51

The RB51 brucellosis vaccine

Strain 19

The Strain 19 brucellosis vaccine

1. Introduction

Brucellosis is a zoonotic disease with worldwide distribution which infects humans in occupational and recreational settings, or via foodborne exposure (CDC ²⁰¹⁷). Of human brucellosis cases reported to the Centers for Disease Control and Prevention (CDC), approximately 70 to 75% occur secondary to consumption of unpasteurised (‘raw’) dairy products from brucellosis‐endemic countries (CDC ²⁰¹⁷).

The RB51 vaccine used in livestock is modified live and an occupational hazard (Ashford et al. 2004; CDC ²⁰¹⁷; Cossaboom et al. ²⁰¹⁸). The RB51 strain of brucellosis is resistant to penicillin and rifampin (CDC ²⁰¹⁷; Schurig et al. ¹⁹⁹¹). In addition, RB51 does not generate a serologic response detectable by extant brucellosis assays, rendering the diagnosis of both human and bovine infections challenging (Ashford et al. 2004; CDC ²⁰¹⁷; Cossaboom et al. ²⁰¹⁸).

RB51 vaccinated animals may become persistently infected with the vaccine strain (CDC ¹⁹⁹⁸; USDA ²⁰¹⁸; Uzal et al. ²⁰⁰⁰) and shed RB51 in their milk (Uzal et al. 2000). USDA recommends pasteurizing all milk products from vaccinated animals (USDA ²⁰¹⁸).

The importance of pasteurisation, and the zoonotic potential of raw milk products produced by RB51‐shedding animals, has been emphasised in a series of human illnesses: a Texas individual who contracted RB51 following consumption of raw milk (Cossaboom et al. 2018), a New Jersey individual with a similar history but unsuccessful initial milk source tracing (Gruber et al. 2020), and a New York individual, the investigation of whom led to identification of the cow responsible for infecting both the NJ and NY patients (Gruber et al. 2020). In this report we describe six RB51 vaccinated cows shedding RB51 in milk, which highlight the importance of understanding RB51 vaccination risks.

2. Materials and Methods

Records from USDA's National Veterinary Services Laboratories (NVSL) database were reviewed. Twenty‐five bovine RB51 cases were identified in the U.S. between 2014 and 2020 based on samples submitted for culture, with the primary complaint being abortion. All U.S. Brucella RB51 cases prior to 2021 detected by NVSL in cattle where the vaccine strain was found in milk were included.

2.1. Culture and Identification at NVSL

Isolation of Brucella bacteria was performed by conventional methods at NVSL (Alton et al. 1988). Bacterial growth appearing typical of Brucella was heat inactivated and identified by modified Brucella AMOS PCR (Bricker and Halling 1995; Bricker and Halling ¹⁹⁹⁴).

2.2. Culture at CDC

After the cow from PA was transported to CDC, milk cultures and susceptibility testing were performed on samples from each quarter within 5 h of collection.

2.3. PCR at CDC

Isolates of some cases (Case 1, 2) were identified via the Bruce‐Ladder PCR at the CDC.

2.4. Whole Genome Sequencing (WGS)

Single colonies were selected for WGS by the CDC to better characterise SNPs within each colony. DNA was extracted using the DNeasy Blood and Tissue Kit (Qiagen, Germantown, MD) from the heat inactivated preparations, and libraries were prepared using the Nextera XT DNA Kit (Illumina, San Diego, CA). Samples were sequenced on a MiSeq (Illumina, San Diego, CA) using 2 × 250 paired‐end reagents and targeting 100× depth of coverage.

2.5. Genome Alignment and Phylogenetic Analysis

WGS obtained from both animal and public health investigations were analysed using NVSL's in‐house vSNP pipeline (see https://github.com/USDA‐VS/vSNP), using the quality parameters outlined in https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4696693/. vSNP is a reference‐based, two‐step pipeline that is an ISO 17025 approved method at the NVSL. Briefly, in step one, sequences were aligned to B. abortus reference genome B. abortus 9–941 (NCBI Accession numbers NC_006932 and NC_006933) using Burrows‐Wheeler Aligner (BWA) (Li and Durbin 2009). Following alignment to the reference, SNPs were identified using Freebayes (Garrison and Marth 2012). Identified SNPs in the form of a variant call format (vcf) file were added to a database of all vcf files. Step two flags and removes unreliable and low‐quality variant calls. Sequences are grouped into user defined clades according to relatedness by identifying common SNPs. For each user defined group, step two output SNP tables in Excel, an aligned FASTA file, and phylogenetic trees constructed with RaxML (Stamatakis 2014) using the aligned whole‐genome SNP sequences under a GTR‐CAT model of substitution and a maximum‐likelihood algorithm. The annotated and position‐referenced SNP tables allowed for quick error identification and correction. SNPs were verified in Integrated Genomics Viewer (IGV) (Robinson et al. 2011; Thorvaldsdóttir et al. ²⁰¹³).

3. Results: Case Reports

3.1. Cases 1 and 2: Texas

In July 2017, the first human case of RB51 linked to consumption of domestically acquired raw milk was reported in the U.S. (Cossaboom 2018; Cossaboom et al. ²⁰¹⁸).

Two Jersey cows in the Texas dairy herd were identified (Case 1 and Case 2) as milk culture positive for strain RB51, and WGS confirmed that the RB51 isolates found in the bulk tank were the most likely source of exposure for the human case (Cossaboom 2018; Cossaboom et al. ²⁰¹⁸).

Investigators euthanised and necropsied the RB51‐shedding cows and submitted tissue samples for culture in September 2017. Four of 20 tissue samples collected from Case 1 tested positive for RB51, and 15 of 20 for Case 2 (Table 1). Case 2 was bacteremic at the time of sample collection, in addition to shedding RB51 in the milk. WGS performed on the isolates from each mammary quarter identified a sequence from the high‐shedding cow (Case 2) as the match for the RB51 found in the bulk tank milk sample and the human patient (Table S1). The dairy returned to selling raw milk in October 2017 (Cossaboom 2018; Cossaboom et al. ²⁰¹⁸).

TABLE 1.

Summary of six Jersey cow cases shedding Brucella abortus RB51 in their milk. All samples were tested at the National Veterinary Service Laboratories in Ames, IA.

Case	1	2	3	4	5	6
State	Texas	Texas	Nebraska	Pennsylvania	Washington	Colorado
Initial sample collection date	9/2017	9/2017	10/2018	12/2018	7/2019	10/2020
Age (mo) at RB51 vaccination	~5	5	11	11	4–6	4–6
Age at detection (year)	3	3	6	8	7	5
Tissues cultured a
Blood	−	+	NS	NS	NS	NS
Milk	+	+	+	+	+	+
Mammary gland	C	+	NS	+	+	NS
Supramammary lymph node	C	+	NS	+	+	NS
Other lymph node(s)	+	+	NS	+	+	NS
Uterus	C	+	NS	−	+	NS
Ovary	C	−	NS	−	NS	NS
Spleen	C	+	NS	−	+	NS
Kidney	C	+	NS	−	NS	NS
Liver	−	−	NS	−	NS	NS
Tonsil	C	C	NS	−	NS	NS
Placenta	NS	NS	NS	NS	+	NS
Fetus	NS	NS	NS	NS	−	NS

^a + = positive, − = negative (no isolation made), C = contaminated, NS = not sampled/submitted.

3.2. Case 3: Nebraska

In October 2018, investigators isolated strain RB51 from a routine milk sample collected at the time of calving from a seven‐year‐old Jersey cow from Nebraska (WGS in Table S1). The cow had been vaccinated with RB51 at 11 months old in 2013. The cow was euthanised due to strain RB51 infection shortly after diagnosis, and no additional samples were collected. This herd's milk was processed for cheese production; no human cases were detected in association with this dairy.

3.3. Case 4: Pennsylvania

In December 2018, investigators tested cows at a raw milk dairy in Pennsylvania after finding a human strain RB51 case in New York. A Jersey cow from the herd was shedding RB51 in her milk. Investigators isolated strain RB51 from all quarters over a 6‐month period, at levels ranging from not detected to 5000 CFU/mL. Investigators sequenced the isolates; two genotypes were detected, each shed from different portions of the udder. The sequence of the isolate obtained from the right front and right rear quarters matched the WGS of the previously undetermined RB51 source of the New Jersey human case diagnosed in October of 2017, while the sequence of the isolate from the left front and left rear quarters matched that of the New York human case diagnosed in November of 2018 (Table S1) (CDC ²⁰¹⁹; Gruber et al. ²⁰²⁰). This RB51‐shedding Jersey cow had been part of the organic raw milk dairy since 2016; the dairy distributed milk to 19 states from 2016 to2018 (Gruber et al. 2020). The cow had been vaccinated with RB51 in 2011, at 11 months old. The affected Jersey cow was purchased, removed from the milking herd, and transported to CDC in Atlanta, GA. In December 2019, the cow was euthanised and necropsied, and samples were sent to NVSL. Brucella strain RB51 was detected in all 4 quarters and the mammary lymph nodes. Interestingly, RB51 isolates (rough strain) from the cow's right front and right rear quarters were inhibited by the presence of rifampicin at 250 μg/mL in media. Broth microdilution testing was conducted at CDC on both the human RB51 isolates and RB51 strains isolated from Case 4's milk, including isolates exhibiting rifampin susceptibility. Though rifampin cut‐offs have not been formally established for Brucella species, the ‘susceptible’ RB51 isolates from the NJ human case and the strains isolated from the right udder of Case 4 exhibited low rifampin MICs (2), whereas the RB51 isolate from the NY human case and strains from the left udder of Case 4exhibited high rifampin MICs (> 8).

3.4. Case 5: Washington

In July 2019, investigators isolated strain RB51 from the aborted placenta of a five‐year‐old Jersey cow in Washington, who had been vaccinated between 4 and 6 months old. The cow was necropsied on‐farm. NVSL isolated strain RB51 from multiple tissues (Table 1). The WGS of isolates from tissue and milk samples was identical (Table S1). The dairy pasteurised their milk; no associated human cases were detected.

3.5. Case 6: Colorado

In October 2020, investigators detected a Jersey cow shedding strain RB51 in Colorado, at a raw milk dairy. NVSL isolated strain RB51 from all four quarters of this cow. Like Case 4, the WGS indicated that only the two left quarters' isolates were the same sequence (Table S1). The three cows that were cohort animals for this cow were tested twice each for brucellosis via serology, milk culture, and milk PCR, all at NVSL, and were negative. No human cases were reported associated with this cow.

4. Discussion

These cases highlight the consequences of vaccinating dairy heifers with the RB51 vaccine and serve as a basis for discussion of its appropriate use, particularly for cows in raw milk operations. In the US, the cases provide evidence to reexamine state‐mandated RB51 vaccination import requirements for cattle in areas outside the GYA (USAHA ²⁰¹⁸), and revise regulations surrounding raw milk sales. In all regions of the world, vaccination of livestock should be undertaken only after careful consideration of the epidemiologic risks and benefits, and in accordance with guidance documents appropriate to the locale (WOAH Manual of Diagnostic Tests and Vaccines for Terrestrial Animals ²⁰¹⁸).

It is worth noting that no additional human cases have been detected in association with the cows discussed above, despite many people being exposed to potentially contaminated milk. The explanation for additional cases not having been detected is likely multifactorial and may be a result of intermittent shedding, dilutional effect, failure to suspect or detect human Brucella cases (which may be a significant issue in certain regions of the world: see Blasco et al. ²⁰²³), or other factors. For Cases 3 and 5, milk was pasteurised. It is also worth noting that the cases described were detected using routine CDC and NVSL methods, which are not optimised for detection of RB51. Regardless of the rarity of human case detection, it is a serious public health concern.

Overall, vaccination of heifers with the RB51 strain poses a significant public health risk to dairy workers and consumers and is a potential cause of human cases that are uniquely difficult to diagnose.

5. Conclusion

The six Jersey cows described shed strain RB51 in their milk following appropriate calfhood vaccination, resulting in at least three human RB51 infections. Given the human and cow cases providing evidence of RB51 as a foodborne illness risk in raw milk, the lack of endemic pasteurisation in various parts of the world, and various diagnostic health capacities of different regions, localities should reevaluate their recommendations and regulations surrounding RB51 vaccination to ensure they are epidemiologically appropriate and responsible.

Funding

The authors have nothing to report.

Ethics Statement

This study was conducted using data collected by state and federal animal health officials and did not require IACUC or IRB approval.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Table S1: Single nucleotide polymorphism table for Brucella abortus RB51 isolates from 6 Jersey cows shedding the organism in their milk.

Hines, E. S. , Margrey S. F., Cox S. W., et al. 2026. “ Brucella abortus RB51 Persistent Infection and Milk Shedding in Vaccinated Dairy Cattle: A Public Health Risk.” Zoonoses and Public Health 73, no. 7: 601–605. 10.1111/zph.70079.

Data Availability Statement

Data available on request due to privacy/ethical concerns.