Study design

The study design was cross sectional. Human patients diagnosed with TBLN at the Felegehiwot Comprehensive Specialized Hospital (located in Bahir Dar City) and their cattle traced to the village of origin of the patients were tested for symptoms of TB, microbial culture evidence of the M. tuberculosis complex species, and lineage using a genetic test. In parallel, a comparable number of cattle owned by TB free households, who live in the proximity of the TBLN patients, were examined. TBLN patients were human patients with enlarged lymph nodes, and who were clinically and cytologically diagnosed as TBLN. TB free households were defined as follows: absence of TB suggestive clinical signs and absence of symptoms at the time of meeting with the investigator collecting data from the consenting individuals. Clinical signs and symptoms included a history of fever and/or cough of greater than two weeks of duration, a failure to gain weight, a loss of appetite, a decline in weight, symptoms of extra-pulmonary TB (EPTB) such as swollen lymph nodes, and the absence of confirmed TB cases in any member of the household during the last 10 years.

Sample and data collection from human subjects

Basic demographic data, and information related to awareness on BTB and its public health implication, and consumption habit of milk and meat were collected from each TBLN patient through an interview using semi-structured questionnaire. Collection of fine needle aspiration (FNA) sample and FNA cytology (FNAC) was done by the pathologist. FNAC was used for the diagnosis of mycobacterial lymphadenitis. FNAC positive samples were drained in a tube containing 1 ml phosphate buffer saline (PBS) solution and used for mycobacterial culture.

Comparative intradermal tuberculin

CIDT was carried out to test cattle for BTB according to the OIE protocol [24]. After two sites, 12 cm apart, on skin of middle third of the neck were shaved and the thickness of each was measured with a caliper, two types of purified protein derivative (PPDs) (supplied by Prionics Lelystad B. V., The Netherlands) were injected intradermally into the two sites. One site was injected with an aliquot of 0.1 ml (ml) of 2500 IU/ml bovine-PPD (B-PPD) and the other was injected with 0.1 ml of 2500 IU/ml Avian-PPD (A-PPD). The skin thickness at each injection site was measured again after 72 h. An animal was considered to be positive for BTB if the skin reaction at the PPD-B site minus the skin reaction at the PPD-A is ≥4 mm.

Culture and spoligotyping

TB cultures were performed using the procedure described by the National TB and Leprosy Control Programme Guideline [25] that was adopted from WHO guideline [26]. Briefly, FNA samples collected from TBLN patients in this study were processed and inoculated on duplicate Lowenstein–Jensen (LJ) slants, one supplemented with pyruvate and the other with glycerol. All the tubes were incubated at 37 °C and slants with no growth at week 8 were considered negative. Bacterial colonies from culture-positive samples were Ziehl-Neelsen stained to identify acid fast bacilli (AFB). Cultures positive for AFB were inactivated by heating at 85 °C for 45 min in a water bath and spoligotyped as previously described [20].

Data analysis

All the statistical data were analyzed by STATA statistical software, version 12 (Stata Corp., Collage station, TX, USA). Chi square test, bivariate and multivariable logistic regression analysis were applied for selected demographic factors verses awareness, and milk and meat consumption habits. Statistical significance was assumed if the confidence interval (CI) did not include one among its values or whenever P value was less than 5%.

The generated spoligotype data were converted into binary and octal formats and entered into the open source spoligotype database available at the website http://www.pasteur-guadeloupe.fr:8081/SITVIT_ONLINE/tools.jsp to establish the lineage, sublineage, and the shared international spoligotype (SIT) number. In addition, the online tool “Run TB-Lineage” (http://tbinsight.cs.rpi.edu/run_tb_lineage.html) was used to predict the major lineages to which the strains belong by a conformal Bayesian network (CBN) analysis.

Level of awareness of the farmers and their food consumption habits

Among the 70 human TBLN patients interviewed, 65.7% (46/70) of the individuals did not know that cattle can transmit BTB (it stands for bovine tuberculosis caused by M. bovis in cattle and other mammals including man). The majority (67.1%, 47/70) of the patients had the habit of consuming raw milk. 54.3% (38/70) of the patients ate only cooked meat products although the cooking temperature levels and time could not be specified. The majority of the patients (73.7%, 28/38) ate cooked meat without understanding the risk of contracting BTB from raw meat.

Tuberculosis in farmers

Tuberculosis in cattle

All the study cattle were tested for BTB with CIDT and interpreted at a ≥4 mm cutoff point. None of the animals were positive for BTB. Further analyses such as microbial cultures and genotypic tests were not conducted to isolate and characterize mycobacteria from cattle tissues. As a result, comparisons with MTBC strains identified in human subjects were not performed.

Transmission of mycobacteria between farmers and their cattle

The results of this study did not provide evidence of direct transmission of the MTBC species between farmers and their cattle in the smallholder farms of northwestern Ethiopia. However, there is preliminary evidence of zoonotic transmission of M. bovis between animals and two human TBLN patients. Further investigations are required to determine whether M. bovis is directly transmitted from animals to their owners.

Level of awareness of the farmers and their food consumption habits

This study evaluates the awareness of respondents as it pertains to BTB and the risk of zoonotic transmission from raw milk and meat consumption habits in relation to their age, sex and educational status. Respondents’ awareness on BTB was generally poor (34.3%, 24/70), and magnifies the public health implication of the disease in the study area. Our finding is consistent with 35, 25.7, 6.9, 29.7 and 15% awareness levels reported in earlier Ethiopian studies [12, 13, 17, 27, 28]. The low level of awareness observed in the present study could also be related to the low level of BTB in the study area. Earlier epidemiological studies have also indicated that the level of disease awareness among famers is related to the prevalence of the disease in that specific area [29] in such a way that awareness on BTB was lower in low prevalent settings compared to high BTB prevalent areas [30].

Awareness of BTB among human TBLN patients in this study was associated with age. Clusters of ages between 29–39 and 50+ years were less likely to be aware of BTB compared to the younger ages (18–29 years). The observed difference in the level of awareness between the different age categories was difficult to interpret. However, it could be related to the expanding nature of education to village level in Ethiopia. As a result, the younger ages have rather more access to education, and had knowledge on BTB and its zoonotic importance compared to the elder one.

The majority (67.1%, 47/70) of the respondents in this study consumed raw milk, and is consistent with the previous study conducted in Central Ethiopia [28], reported 79.3% (46/58) of livestock holders had habit of raw milk consumption. We, therefore, suggest that improving knowledge and awareness of milk borne transmission is important to prevent zoonotic TB in human in the study area.

Tuberculosis in farmers

Molecular characterization by spoligotyping identified strains of M. tuberculosis as the major causative agents of TBLN in human patients participated in this study. This finding is compatible with previous studies in Ethiopia [31–33]. However, M. bovis was isolated from the two TBLN cases even though cattle owned by these patients were none reactor for CIDT test. Thus, we suggest the potential zoonotic transmission of M. bovis from animal since the two TBLN cases with M. bovis were identified by the questionnaire as consumers of raw milk and undercooked meat product.

SIT53 and SIT149 were the dominant strains identified in TBLN patients participated in the present study. These strains also isolated in human FNA and sputum samples collected and studied previously in Ethiopia [32, 34–36], and suggest the presence of similarity in the population of mycobacterial strains that causes PTB and EPTB.

Tuberculosis in cattle

We did not demonstrate that TB is transmitted from human TBLN patients to their cattle in this study. All cattle were found negative for BTB after CIDT test, and none of the cattle tissue had cultured for mycobacterial growth. However, SIT53 and SIT149, which were isolated from human TBLN patients in the present study, were also isolated from cattle [21, 23] and goat [22] previously in Ethiopia, which indicate possible human-to-animal transmission.

Transmission of mycobacteria between farmers and their cattle

A potential transmission of zoonotic TB was observed in this study since there was isolation of M. bovis from two human TBLN cases even though direct link of transmission between specific cattle and its owners were not confirmed. Similar findings were reported from previous Ethiopian studies [9, 19, 20]. These studies isolated M. bovis from human pulmonary TB (PTB, it is an infectious disease principally caused by M. tuberculosis, which is primarily a human pathogen and characterized by the growth of tubercles in the tissues, especially the lungs) and/or TBLN cases, and suggested the occurrence of transmission of zoonotic TB between livestock and humans. However, despite close contact between humans and livestock, and low level of awareness of zoonotic TB, including the high consumption rate of raw milk by farmers in the study area, the prevalence of M. bovis in human TBLN patients recorded in the present study was lower than expected. This could be related to the overall low prevalence rate of BTB in the study area as it was shown in the present study and 1.27% prevalence reported previously in northwestern Ethiopia [18]. The small sample size (70 TBLN) could also contribute to the recorded lower isolation rate of M. bovis from farmers in the study area.