Neglected zoonotic agents in cattle abortion: tackling the difficult to grow bacteria

Background Coxiella burnetii, Chlamydia abortus and Leptospira spp. are difficult to grow bacteria that play a role in bovine abortion, but their diagnosis is hampered by their obligate intracellular lifestyle (C. burnetii, C. abortus) or their lability (Leptospira spp.). Their importance is based on the contagious spread in food-producing animals, but also as zoonotic agents. In Switzerland, first-line routine bacteriological diagnostics in cattle abortions is regulated by national law and includes only basic screening by staining for C. burnetii due to the high costs associated with extended spectrum analysis. The aim of this study was to assess the true occurrence of these zoonotic pathogens in 249 cases of bovine abortion in Switzerland by serology (ELISA for anti-C. burnetii and C. abortus antibodies and microscopic agglutination test for anti-Leptospira spp. antibodies), molecular methods (real-time PCR and sequencing of PCR products of Chlamydiales-positive cases), Stamp’s modification of the Ziehl-Neelsen (mod-ZN) stain and, upon availability of material, by histology and immunohistochemistry (IHC). Results After seroanalysis the prevalence was 15.9% for C. burnetii, 38.5% for C. abortus and 21.4% for Leptospira spp. By real-time PCR 12.1% and 16.9% of the cases were positive for C. burnetii and Chlamydiales, respectively, but only 2.4% were positive for C. burnetii or Chlamydiales by mod-ZN stain. Sequencing of PCR products of Chlamydiales-positive cases revealed C. abortus in 10% of cases and the presence of a mix of Chlamydiales-related bacteria in 5.2% of cases. Pathogenic Leptospira spp. were detected in 5.6% of cases. Inflammatory lesions were present histologically in all available samples which were real-time PCR-positive for Chlamydiales and Leptospira spp. One of 12 real-time PCR-positive cases for C. burnetii was devoid of histological lesions. None of the pathogens could be detected by IHC. Conclusion Molecular detection by real-time PCR complemented by histopathological analysis is recommended to improve definitive diagnosis of bovine abortion cases and determine a more accurate prevalence of these zoonotic pathogens. Electronic supplementary material The online version of this article (10.1186/s12917-017-1294-y) contains supplementary material, which is available to authorized users.


Background
Abortion in dairy cattle is one of the major causes of economic loss in the livestock industry [1] and three of the bacterial agents that are implicated in bovine abortion during mid-to late-gestation are the difficult to grow: Coxiella burnetii, Chlamydia abortus and pathogenic Leptospira spp. Their importance is based on not only in the economic loss in animal production but also in their zoonotic risk [2][3][4].
C. abortus and C. burnetii are obligate intracellular Gram-negative bacteria. C. abortus, the causative agent of ovine enzootic abortion, may also lead to reproductive disorders in large ruminants [2,5] and is known to cause spontaneous abortion in pregnant women [5,6]. Other members of the families Chlamydiaceae, Parachlamydiaceae and Waddliaceae have also been found to play a possible role in abortion in ruminants as well as in humans [7][8][9][10][11]. C. burnetii has a wide host range, including domestic and wild animals. Infection in most animals is subclinical or presents with non-specific clinical signs, whereas ruminants, the main reservoir of infection, may present with late abortion and stillbirths; moreover, C. burnetii might be associated with metritis and infertility in cattle [3,[12][13][14][15][16][17][18]. Human infection with C. burnetii is known as Q fever and can lead to miscarriage in women [19][20][21]. Leptospirosis is caused by Gram-negative, pathogenic spirochetes of the genus Leptospira that is divided in more than 250 pathogenic serovars worldwide, which are classified into 25 serogroups on the basis of their serological phenotype. In cattle, leptospirosis is mainly associated with reproductive problems including infertility, low conception rate, abortion, stillbirths and weak offspring [22][23][24]. Cattle are considered to be the maintenance host of serovar Hardjo resulting in a high degree of subclinical infections [25]. Human leptospirosis occurs worldwide, is transmitted via direct or indirect contact with urine from infected animals and is due mostly to recreational and occupational activities [26][27][28]. Numerous outbreaks of leptospirosis worldwide have been also associated with heavy rainfall and flooding [29,30]. Abortion in women due to leptospirosis may occur if infection takes place during pregnancy [31,32].
Given the numerous possible etiologies of abortion in ruminants and the high cost of definitive diagnosis, a finance-limited investigation is performed usually, and the causative agent often remains undetermined [33]. Of the three cattle abortifacient pathogens discussed, only investigation of C. burnetii is legally regulated in Switzerland requiring Stamp's modification of the Ziehl-Neelsen (mod-ZN) stain [34] of tissue smears [Ordinance on Epizootic Diseases (TSV) SR.916.401; Article 129].
According to the epizootics database of the Swiss Federal Food Safety and Veterinary Office (InfoSM www.infosm.blv.admin.ch, consulted on 02/08/2017), 676 cases of coxiellosis in cattle were reported from 2006 to 2016. Although pathogenic Leptospira spp. and C. abortus are not included in routine bovine abortion diagnostics and, moreover, Leptospira interrogans serovar Hardjo is exempt from mandatory notification in cattle, 43 cases of leptospirosis and 23 cases of chlamydiosis in cattle were reported in the same time frame.
In this study, the recommended mod-ZN method was complemented with serology, molecular methods, histology and immunohistochemistry to determine the degree of underestimation of the three abortifacient pathogens C. burnetii, C. abortus and pathogenic Leptospira spp. in bovine abortion in Switzerland.

Stamp's modification of the Ziehl-Neelsen stain
Smears of placentas, abomasal contents (n = 299) and the negative control placenta were subjected to mod-ZN staining [32] and examined by light microscopy. Chlamydia-positive placental tissue was included as positive control in every stain. The sample was considered positive for Chlamydiales and/or C. burnetii when intracytoplasmic red-stained coccobacilli appeared in clumps against a blue background. The technique does not allow a differentiation between Chlamydiales and C. burnetii.

Serological studies
The 182 maternal sera were tested for antibodies against C. burnetii and C. abortus using the commercial CHEKIT® Q fever antibody ELISA Test Kit and CHEKIT® C. abortus Antibody Test Kit (IDEXX, Liebefeld-Bern, Switzerland) according to the manufacturer's instructions. The results were expressed as S/P values and derived from the ratio between optical density (OD) of the sample (S) and the OD of positive control (P) included in the kits. IDEXX state an S/P ≥ 40% is considered positive, an S/P < 30% is considered negative, and S/P values between these are considered suspect positive.
The serological detection of antibodies against Leptospira spp. was performed by microscopic agglutination test (MAT) (Manual of Diagnostic Tests and Vaccines for Terrestrial Animals of the Ordinance of Epizootic Diseases [22]). Twelve serovars were included in the test panel: Australis, Autumnalis, Ballum, Bataviae, Bratislava, Canicola, Grippotyphosa, Hardjo, Icterohaemorrhagiae, Pomona, Sejroe and Tarassovi (Additional file 1: Table  S1). Sera were screened initially for agglutination at a dilution of 1:100 in sterile 0.85% NaCl. Reactive sera were titrated in two-fold serial dilutions to determine the end-point titer defined as the dilution at which at least 50% agglutination occurs. In every serological analysis negative and positive control sera were included as controls.

DNA extraction and molecular studies
For the extraction of total genomic DNA 2 g of placenta or 2 mL of fetal abomasal content were suspended in 5 mL 0.85% NaCl in an IKA® DT-20 tube [35] and homogenized twice for 30 s at 6000 rpm, using the IKA ULTRA-TURRAX® tube drive. Subsequently, 500 μL of the homogenates were used for DNA extraction using QIAamp Mini Kit (Qiagen, Hombrechtikon, Switzerland). Fluorometric quantification of DNA was performed by Quantus™ Fluorometer (Promega, Dübendorf, Switzerland).
Real-time PCR targeting the lipL32 gene of Leptospira spp. was performed using primers and probe described by Villumsen et al. [38]: LipL32-F: 5′ AGAGGTCTT-TACAGAATTTCTTTCACTACCT 3′; LipL32-R: 5′ TGGGAAAAGCAGACCAACAGA 3′; probe LipL32-P: 5' 6FAM-AAGTGAAAGGATCTTTCGTTGC-MGBN FQ 3'. PCR was performed in a total volume of 25 μL, 1X final concentration of TaqMan Universal PCR Master Mix, 1 μM of each primer, 80 nM of the probe, 0.5X of IPC Template and 0.5X IPC Mix and 2.5 μL of the template. The following conditions were applied: 94°C for 2 min, 45 cycles of 94°C for 15 s and 60°C for 30 s using the TaqMan 7500 Fast Real-time PCR System. DNA of Leptospira spp. serovar Icterohaemorrhagiae strain RGA and water were used as positive and negative controls, respectively. Samples were considered positive when showing an exponential amplification curve up to cycle 40 in both replicates.

Histopathology
To assess the significance of the molecular analysis, all cases with real-time PCR-positive results were examined histopathologically and by IHC (n = 32) when the placental tissue was available and was not severely autolytic. Selected samples of placenta were fixed in buffered formalin (10%), processed routinely through graded alcohols and embedded in paraffin-wax. Sections (4 μm) were mounted on Thermo Scientific™ SuperFrost Plus© (Braunschweig, Germany) slides and stained with hematoxylin and eosin (HE) for histological evaluation.
Briefly, for all three antibodies 4 μm thick sections were deparaffinized and rehydrated through graded alcohols.
For C. burnetii, sections were immersed in 3% H 2 O 2 in methanol (v/v) for 20 min to quench endogenous tissue peroxidases. Non-specific antibody binding was blocked with 25% normal goat serum (NGS, Vector Laboratories, Peterborough, UK) in TBS for 30 min and incubated with the primary antibody overnight at 4°C. Visualization of the bound anti-C. burnetii primary antibody was by EnVision Kit (goat anti-mouse horseradish peroxidase conjugate, DakoCytomation, Ely, UK) according to the manufacturer's instructions followed by addition of the chromogen 3-amino, 9-ethyl-carbazole (AEC, Vector Laboratories, Peterborough, United Kingdom) for 10 min.
Immunohistochemistry for pathogenic Leptospira spp. was performed using the avidin-biotin-peroxidase complex (ABC) method. Sections were treated with 0.5% H 2 O 2 in methanol (v/v) for 30 min to block endogenous peroxidase, heated in sodium-citrate buffer for 30 min in the microwave for antigen retrieval, incubated with 20% goat serum for 30 min, then incubated with the respective primary antibody overnight at 4°C. Biotinylated goat-anti-rabbit IgG (BA-1000) diluted 1:200 in PBS (Vector Laboratories, Burlingame, CA, USA) was used as secondary antibody with incubation time of 60 min. Colour development was with 3,3′-diaminobenzidine tetrahydrochloride (DAB) with H 2 O 2 (0.03%, pH 7.2) for 5 min.
Sections immunolabeled with the respective primary antibodies against Chlamydiaceae, C. burnetii and pathogenic Leptospira spp. were all counterstained with hematoxylin prior to mounting in an appropriate mountant.
Four qPCR-negative cases for all three agents were included as negative controls. For C. burnetii and pathogenic Leptospira spp., primary antibodies were substituted with an isotype matched normal mouse IgG antibody or normal rabbit IgG (1:3000; R4505; Sigma Aldrich, Taufkirchen, Germany), respectively, as method negative control preparations.
Sections of intestinal tissue from gnotobiotic piglets experimentally infected with porcine Chlamydia suis strain S45/6, C. burnetii-positive sheep and human placentas and hamster kidney infected with L. interrogans serovar Hardjo JB191 were included as positive controls.

Statistical analysis
We calculated the degree of agreement between the serological and the molecular tests for C. burnetii, C. abortus and Leptospira spp. carried out in 182 cases using Cohen's kappa (κ) coefficient with 95% of CIs with the online software GraphPad (http://graphpad.com/ quickcalcs/kappa2). Standard cutoffs were used to define poor (κ < 0.40), fair (κ = 0.41-0.60), good (κ = 0.61-0.80) and very good agreement (κ ≥ 0.80). The techniques that do not allow for detection of a specific pathogen or yielded only negative results were not included in the comparison.

Stamp's modification of the Ziehl-Neelsen stain
Of the 299 tissue smears, 10 placental smears and two of abomasal contents were positive as denoted by the presence of red intracytoplasmic organisms consistent with coccobacilli. One of the positive placenta and abomasal content samples were from the same case (Additional file 2: Table S2).

Detection by serological analysis
Of the 182 sera tested, 29 (15.9%) were positive for C. burnetii and two (1.1%) were suspect positive. Chlamydial antibodies were detected in 70 (38.5%) of the 182 sera and 23 (12.6%) sera were suspect positive. The prevalence of antibodies against Leptospira spp. was 39/182 (21.4%), with 21 (11.5%) sera being positive for at least two serovars. Serovar Hardjo was the most frequent (31/39) followed by serovar Sejroe (14/39). Yet, 12 sera were positive for both serovars with 10 sera showing a higher titer for Hardjo and, hence, indicating that the latter is the causative serovar. Six cases were positive for serovar Australis (Table 1).

Detection by molecular analysis
Real-time PCR detection of C. burnetii was positive in 28/242 (11.6%) placenta and 7/57 (12.3%) abomasal content samples. The pan-Chlamydiales real-time PCR was positive for 41/242 placenta (16.9%) and 2/57 (3.5%) abomasal content samples. The results after amplicon sequencing of positive samples with a Ct ≤ 35 are summarized in Table 2. C. abortus was detected in 24 placenta samples and in one abomasal content sample. New Chlamydia-related bacteria were detected in four placenta samples. All four were distantly related to known species but all four could be assigned to the Parachlamydiaceae family since they exhibited >90% similarity with at least a member of this clade. The sequencing of the remaining 12 samples was not discriminatory because of superposition of sequences, probably due to the presence of more than one member of the Chlamydiales order in the sample. Real-time PCR for the detection of Leptospira spp. was positive for 14/242 (5.8%) samples of placenta and 1/57 (1.8%) samples of abomasal contents. For seven of the samples positive by realtime PCR, serum for analysis by MAT was available and was positive in five cases (serovar Hardjo, n = 4; serovar Icterohaemorrhagiae, n = 1) and negative in two cases. All the results of the molecular analyses are   Table S2 and the percentage of positive placenta and abomasal content samples is summarized in Fig. 1.
To compare serological and molecular techniques we analyzed 182 cases that were processed by ELISA, MAT and PCR (Fig. 2). All three pathogens had more positive results in the serological analyses than in the molecular analysis, with C. abortus having the highest seropositivity. Results from samples that were positive by real-time PCR for more than one pathogen are summarized in Table 3.

Histopathology and IHC
The severity of the placentitis, necrosis, inflammatory cell infiltrate and vasculitis in all real-time PCR-positive cases of C. burnetii, Chlamydiales and Leptospira spp. that were evaluated histologically (when tissue was available and not autolytic) varied greatly and the histological findings are summarized in Table 4.
Regardless of the etiological agent, if necrosis was present in the cotyledon it was multifocal, randomly distributed and affected the chorioallantoic stroma and the villi including the trophoblasts.
Placentitis was present in 12 of 13 samples that were positive for C. burnetii by real-time PCR. Necrosis was present in 11 of 13 cases. Mixed inflammatory infiltrates characterized by neutrophils, macrophages and lymphocytes were present in nine of 13 cases and only three cases were designated as suppurative placentitis. Vasculitis was present in five of 13 cases and characterized by infiltration of neutrophils, macrophages and lymphocytes primarily in the tunicae media and adventitia resulting in mild fibrinoid necrosis only (Fig. 3a). All cases positive for Chlamydiales by real-time PCR had placentitis and necrosis. Mixed inflammatory cell infiltrates were present in 12 of 14 cases, only one was designated suppurative and one case was infiltrated by macrophages and lymphocytes (mononuclear) only. Vasculitis was present in seven of the 14 cases (Fig. 3b). Similarly, all cases of Leptospira spp. positive by real-time PCR had placentitis and necrosis (Fig. 3c). Mixed inflammatory cell infiltrates were present in three of five cases, while in two cases the inflammation was comprised of mononuclear leukocytes only. For cases positive for Leptospira spp. by real-time PCR no suppurative placentitis was found and only one case had vasculitis.
Immunohistochemical analysis failed to visualize C. burnetii, Chlamydiales and Leptospira spp. in all sections evaluated (Table 4).

Statistical analysis
The Cohen's kappa coefficient is summarized in Table 5. The degree of agreement between the serological and the molecular diagnostic techniques for C. burnetii, C. abortus and Leptospira spp. was poor with κ = 0.103 ± 0.086, κ = −0.006 ± 0.067 and κ = 0.163 ± 0.074, respectively.

Discussion
This study investigated, concomitantly, the prevalence of C. burnetii, C. abortus and pathogenic Leptospira spp. DNA in bovine abortion material and seroconversion in affected dams and highlights the underestimation associated with using a single staining technique. Although detection of any of these pathogens does not equate invariably to causality with respect to bovine abortion, their presence does invariably represent a high zoonotic risk and a possible reservoir of infection for other animals.
The frequency of antibodies specific for C. burnetii in dams was 15.9%, comparable to the reported seroprevalence of 16.7% in aborting cows in Switzerland by Hässig Fig. 1 Incidence of placenta and abomasal content samples that were positive by real-time PCR for Coxiella burnetii, Chlamydiales and/or pathogenic Leptospira spp. and Lubsen [40]. However, previous studies reported similar seropositivity for C. burnetii antibodies but in healthy cows and in different European countries (Bulgaria 20.8%; France 15%; Germany 19.3% and the Netherlands 21%) [41] suggesting serological results are not reliable for diagnostic purposes. With 38.4% positive and 13.0% suspect positive reactions the seropositivity to anti-C. abortus antibodies was the highest of the three abortifacient pathogens investigated in this study. This result was in agreement with studies in other countries which also reported a high prevalence of anti-chlamydial antibodies in cattle, with seropositivity ranging from 45% to 100% [42]. However, such high seropositivity rates have to be interpreted with caution. Firstly, a single seropositive result is not necessarily related to the etiology of the abortion and might be due to a previous exposure [43]. Secondly, serological tests may not be C. abortus specific and positive titers can arise from crossreactivity to C. pecorum, a common intestinal opportunistic chlamydial species found in ruminants [44]. The frequency of Leptospira spp. antibodies was 21.4% and primarily due to serovar Hardjo (seroprevalence of 17.0%). The implication of a positive titer to serovar Hardjo on fetal loss remains controversial as many studies failed to show a causal association between seropositivity and abortion [22,[45][46][47][48], while others described Hardjo as a cause of abortions [49,50]. The interpretation of the serological results for Leptospira spp. should be done carefully since there is cross-reactivity between serovars of the same serogroup; nevertheless, the infecting serovar is more likely to show the higher titer [22,51]. Although Sejroe was the second most prevalent serovar (14 cases), 12 cases were positive for serovars Hardjo and Sejroe, belonging to the same serogroup. Yet, Hardjo presented the higher titer in 10 cases making Australis (six cases) the second most frequent serovar. Abortion in cattle due to serovar Hardjo is a chronic event with a variable serological response at the time of abortion [22] and confirmation of infection by MAT is difficult because maternal antibody production mostly occurs prior to fetal death [52].
Molecular detection of DNA of abortifacient agents has been shown to be highly sensitive and specific [37,[53][54][55]. By real-time PCR we detected C. burnetii in 12.1% of cases, similar to findings in Italy (11.3%) obtained by nested PCR [56] but lower than results by classical PCR from Portugal (17.2%) [57] and by real-time PCR from Hungary (25.9%) [58]. These findings, although obtained with different techniques, may reflect different endemicity. Furthermore, we showed the capacity of C. burnetii of spreading via the amniotic-oral route [59,60] with the 7/57 samples of abomasal content being positive. Of the 21 real-time PCR positive cases of C. burnetii, 15 were seronegative suggesting early stages of infection when antibodies are not yet present, or environmental contamination of samples or failure of the dam to seroconvert occurred. In contrast, 23 cases with positive sera were negative by real-time PCR suggesting previous exposure to C. burnetii is not uncommon. The statistical analysis showed a poor agreement (κ = 0.103 ± 0.086) between the serological and the molecular technique indicating that there is a poor relationship between the seropositivity of the dam for antibodies to C. burnetii and an abortion event as reported previously [61][62][63][64]. It is important to keep in mind that real-time PCR is highly sensitive and thus able to detect low levels of C. burnetii. Yet, different strains harbor a very variable number of the target IS1111 (between 7 and 110) making quantification inaccurate for this bacterium [65]. For the final interpretation at herd-level it is recommended to include complementary techniques and consider the case history [18,66].
Of 43 real-time PCR-positive samples for Chlamydiales (placenta, n = 41; abomasal content, n = 2) C. abortus could be identified by sequencing in 9.6% of the cases, although the prevalence could be higher because in 12 samples a single species could not be assigned due to  multiple peaks. In Eastern Switzerland, C. abortus was considered not to play an important role in bovine abortion in studies by end-point PCR [39] and real-time PCR [67]. However, Blumer et al. [9] confirmed the presence of C. abortus in 14.8% of studied cases of abortion from Eastern Switzerland. We could detect members of the Parachlamydiaceae family in four samples confirming that Chlamydia-related bacteria could be involved in bovine abortion as reported previously [9,39,67] and could cause mixed infection [68]. It is noteworthy that some samples with Chlamydia-related bacteria, including P. acanthamoebae, were positive by C. abortus ELISA also. This result might be due to the production of antibodies that could cross-react with other chlamydialspecies due to a genus-specific epitope of the lipopolysaccharide [44,[69][70][71]. This might also be the underlying reason for the poor agreement (κ = −0.006 ± 0.067) between the serological and the molecular technique. In six cases of coxiellosis we found evidence of coinfection with C. abortus and Chlamydiales-related bacteria. Although C. burnetii and Chlamydiales belong to phylogenetically unrelated species [72], they have some similarities in their interaction with the host and mechanisms of pathogenicity [73]. Thus, the diagnosis of either agent is usually established by microscopic examination of stained placenta smears in veterinary diagnostic laboratories but this cannot discriminate between the different organisms. Pritchard et al. [74] stated that the mod-ZN stain is insufficiently sensitive in cattle cotyledons. Our findings agree with this and confirm that the mod-ZN stain is not very sensitive for the detection of either C. burnetii or Chlamydiales infection in bovine abortion material and that it should be replaced by specific real-time PCRs.
Pathogenic Leptospira spp. had a prevalence of 5.6% by real-time PCR (placenta, n = 14; abomasal content, n = 1). The detection of leptospires in internal organs of aborted or stillborn fetuses reflects chronic leptospirosis of the mother and indicates an active infection of the fetus, but PCR-based diagnosis of leptospirosis alone cannot identify the infecting serovar; moreover, contamination with faeces or autolysis in clinical samples is known to lead to false-negative results [22]. Hence, the combination of both, molecular and serological techniques is of epidemiological value, even though no satisfactory agreement between techniques (κ = 0.163 ± 0.074) was achieved. Unfortunately, in only seven cases material was available for both analysis, and four sera of these were positive for serovar Hardjo and one for serovar Icterohaemorrhagiae. Two of the samples positive by real-time PCR for Leptospira spp. DNA were negative in MAT indicating an early stage of the infection or failure to detect seroconversion. In one case, pathogenic Leptospira spp., presumably identified as serovar Hardjo by serology, were detected together with C. burnetii and Chlamydiales and in another case we found possible coinfection between pathogenic Leptospira spp. and C. abortus.
Histological investigation and confirmation of the cellular inflammatory process indicative of infectious agents is important to unambiguously confirm the implication of a specific etiological agent especially if it could also be present in the commensal and the environmental microbiota [43]. However, as the cotyledonary lesions are not pathognomonic for any of the three pathogens investigated [39,75], a definitive diagnosis based on histopathology only is not possible. Accordingly, in real-time PCR-positive cases of C. burnetii and Chlamydiales we found similar placental lesions varying only in the degree and severity of the inflammatory infiltrate. Although, vasculitis in the placenta of abortion cases is described as a prominent feature of C. abortus infections [39], it is not invariably present. Furthermore, vasculitis in the placenta is present frequently in cases of C. burnetii abortion [75,76] as was observed in this study.  Table 5 Cohen's kappa (κ) coefficient with 95% of CIs to assess the degree of agreement between molecular and serological techniques for the diagnosis of C. burnetii, C. abortus and Leptospira spp.  Additionally, we found that not all Chlamydiales real-time PCR-positive cases displayed vasculitis, similar to previous reports [9,39]. All Leptospira spp. real-time PCR-positive cases showed necrotizing placentitis with three and two displaying mixed and mononuclear inflammatory infiltrates, respectively. Vasculitis was not observed in any sample except one case which was real-time PCR-positive for C. burnetii and Chlamydiales also. Placental lesions caused by Leptospira spp. in bovine abortion are not well characterized but, based on our limited observations, vasculitis is not a prominent feature. Lesion-associated pathogen detection is usually considered vital for definitive diagnosis to prove causality. However, we were not able to identify lesion-associated C. burnetii, Chlamydiales or Leptospira spp. by IHC in any of the analyzed slides. IHC is known to have lower sensitivity than real-time PCR [77,78], especially when there is some degree of autolysis in the samples as is often the case for abortion material.
Limited first-line diagnostics (mod-ZN staining) could only detect possible abortifacient agents in 11 cases (4.4%) while real-time PCR detected a possible abortifacient agent in 78 cases (31.2%). The fact that (i) C. burnetii, Chlamydiales and Leptospira spp. are all difficult to culture, (ii) serology cannot exclude a past infection or confirm an ongoing infection and (iii) IHC apparently fails to demonstrate the presence of the agents, makes the molecular approach the method of choice.

Conclusions
In conclusion, we recommend an extended workflow including molecular analysis for routine abortion diagnostics to avoid the underestimation of the discussed agents and histological analysis to avoid misinterpretation of realtime PCR positive results. It would be prudent to use molecular methods initially and then subject positive cases to histological screening. For further epidemiological investigations complementary serological analyses should be considered. However, the real value of this work was determining the inherent public health risk with respect to these zoonotic pathogens and their prevalence in bovine abortion material as important source of infection.

Additional files
Additional file 1: Table S1.