Infectious Agents Identified by Real-Time PCR, Serology and Bacteriology in Blood and Peritoneal Exudate Samples of Cows Affected by Parietal Fibrinous Peritonitis after Caesarean Section

The aim of this study was to identify the pathogens potentially involved in parietal fibrinous peritonitis (PFP). PFP is a complication of laparotomy in cattle, characterized by an accumulation of exudate inside a fibrinous capsule. We have studied 72 cases of PFP in Belgian blue cows, confirmed by a standard diagnostic protocol. Blood was collected to evaluate the presence of antibodies for Mycoplasma bovis (M. bovis), Coxiella burnetii (C. burnetii) and Bovine Herpesvirus 4 (BoHV4) by enzyme-linked immunosorbent assays. Peritoneal exudate was obtained from the PFP cavity to perform bacteriological culture, and to identify the DNA of M. bovis, C. burnetii and BoHV4 using real time polymerase chain reaction (qPCR). Bacteriological culture was positive in most peritoneal samples (59/72); Trueperella pyogenes (T. pyogenes) (51/72) and Escherichia coli (E. coli) (20/72) were the most frequently identified. For BoHV4, the majority of cows showed positive serology and qPCR (56/72 and 49/72, respectively). Contrariwise, M. bovis (17/72 and 6/72, respectively) and C. burnetii (15/72 and 6/72, respectively) were less frequently detected (p < 0.0001). Our study proves that PFP can no longer be qualified as a sterile inflammation. Moreover, we herein describe the first identification of BoHV4 and C. burnetii in cows affected by PFP.


Introduction
Parietal fibrinous peritonitis (PFP) in cattle is a postoperative complication of laparotomy [1][2][3], characterized by the accumulation of fibrin and peritoneal exudate inside a thick fibrous capsule between the outer sheath of the parietal peritoneum and the abdominal muscular layers [1,3]. Symptoms of PFP occur several weeks after surgery, and may include hyperthermia, anorexia, weight loss, visual abdominal distention, and colic [3][4][5].
In Belgium, PFP is frequently encountered in rural veterinary practice due to the large number of elective caesarean sections (CS) performed in the Belgian Blue breed [6]. Its incidence after CS has been Vet. Sci. 2020, 7, 134; doi:10.3390/vetsci7030134 www.mdpi.com/journal/vetsci estimated to be 1%, and its mortality has been estimated at 13% [7,8]. Unfortunately, PFP is very scarcely documented and rural practitioners have little information considering its treatment, prevention and prognosis. In particular, the aetiology of PFP is the subject of speculation. For a long time, PFP has been considered as an aseptic inflammation [5,8,9]. This assumption has recently been challenged, after the isolation of several aerobic and anaerobic bacteria such as Trueperella pyogenes (T. pyogenes), Escherichia coli (E. coli), Staphylococcus aureus, Pseudomonas aeruginosa, Proteus mirabilis, Fusobacter necrofurum, Comamonas kerstersii, Bacillus licheniformis, and Bacteroides species in in the peritoneal fluid of PFP cows [2][3][4]. The last decade, three infectious agents have received ample attention in Belgian rural practice, i.e., Mycoplasma bovis (M. bovis), Coxiella burnetii (C. burnetii) and Bovine Herpesvirus 4 (BoHV4). All three germs have been identified in various bovine infectious disorders including reproductive tract disease, abortion, mastitis, respiratory diseases, and arthritis [2,[10][11][12], and the number of positive laboratory diagnoses shows an increasing trend [13,14]. Their implication in peritonitis and PFP is unclear.
The aim of this study was to perform bacteriological culture on peritoneal fluid samples of a large cohort of cows presenting PFP. Furthermore, to gain a broader insight into the aetiology of PFP, we aimed to evaluate the implication of M. bovis, C. burnetii and BoHV4 in PFP, by determining the presence of antibodies in the blood and genetic material in peritoneal fluid samples from cows presenting PFP.

Material and Methods
Between March 2017 and March 2018, the Clinical Department of Production Animals (University of Liege in Belgium) and the Regional Association of Animal Health and Identification (ARSIA) collaborated with Belgian rural veterinary practitioners in a project to obtain diagnostic elements on PFP in Belgian blue cattle breed. All rural veterinarians from the ARSIA database were contacted by e-mail, and instructions for the diagnosis and treatment of PFP were published online. In each case suspected of PFP based on clinical signs and ultrasound findings, an aseptical paracentesis of the cavity was carried out to confirm the diagnosis, as described in previous studies [3,4]. Finally, PFP was confirmed in 72 cases, and 10 mL of peritoneal fluid was collected in each case for further examination. Furthermore, blood samples were obtained from the coccygeal vein using non-coagulant Vacutainer ® tubes (BD, Plymouth, UK) for biochemical analysis, and an additional blood tube was collected for further examination of the serological status of the animals. Finally, the treatment (and definitive diagnosis) consisted of surgical draining of the PFP cavity. All blood and peritoneal fluid samples were kept at 4 • C and dispatched to the ARSIA laboratory. The national identification database (SANITEL) was consulted afterwards for the cattle age. Consent was obtained from all veterinarians and owners to use the samples to perform the current study and publish the results.
All invasive procedures (paracentesis, blood sampling and surgical drainage) were done in cases encountered in the field, primarily for diagnostic and therapeutic purposes. At no point did the research protocol interfere with treatment decisions and housing or management of the cows. Therefore, the animals in our study did not fall into the definition of an experimental animal, and no ethical approval was required.
Peritoneal exudate samples were used for aerobic and anaerobic bacteriological culture and for the detection of C. burnetii, M. bovis and BoHV4 genetic material. Blood samples were used for the detection of C. burnetii, M. bovis and BoHV4 antibodies.
The samples for aerobic culture were grown on Columbia agar, Gassner and Columbia/Nalidixic acid agar media (Thermo Fisher Scientific, Brussels, Belgium) at 37 ± 2 • C. Samples for anaerobic culture were grown under anaerobic conditions on Schaedler medium (Thermo Fisher Scientific, Brussels, Belgium) at 37 ± 2 • C. Two readings of each medium were performed at 18 to 24 h and 36 to 48 h of incubation. Bacterial identification was performed by the Maldi Biotyper ® (Bruker Daltonics, Bremen, Germany). The culture was considered "negative" if no bacterial growth was observed, and "positive" when one or several bacteria were found.
The ELISA test results of BoHV4 and C. burnetii are semi-quantitative. The antibody concentration (%) is calculated as the ratio between the optic density of the tested sample and a control sample, multiplied by 100. Results for BoHV4 and C. burnetii were classified as "negative" (relative density below 30% and 40%, respectively), as "positive" (relative density between 30% and 120% and between 40% and 300%, respectively) or as "highly positive" (relative density above 120% and above 300%, respectively). The results for the M. bovis ELISA kit are only qualitative ("positive" or "negative").
Statistical analyses were performed using SAS (2001). Descriptive analysis was carried out for the age of cows and the number of bacteria cultured in the peritoneal exudate. Continuous data (age, number of bacteria found by bacteriology) were checked for normal distribution with a Shapiro-Wilk test, and displayed as the median and range in case of non-normal distribution. Chi-square and Fisher tests were used to compare the number of positive and negative results of bacteriological culture, ELISA and qPCR, and to compare the germ-specific proportions within positive samples of bacteriology, ELISA and qPCR. Moreover, a Chi-square test was used to compare the antibody and DNA concentration in the positive, semiquantitative samples. A test of independence was performed to evaluate the relation between ELISA and qPCR outcomes and the results of bacterial culture. The procedure "Proc Freq" in SAS was used for all statistical analyses; the cut-off of significance was fixed at p < 0.05.

Results
In total, 37 rural veterinarians collected blood and peritoneal exudate samples from 72 cows affected by PFP after CS in 61 Walloon farms. The age of cows affected by PFP varied from 26 to 120 months with a median of 45 months.
Bacteriology was positive in the majority (59/72) of cows and negative in only 13/72 samples (p < 0.0001). The number of bacteria identified in the positives samples varied between 1 to 3 with a median of 1.
In total, 82 different strains from nine bacteria species were identified in the positive samples, among which T. pyogenes (51/59) and E. coli (20/59) were predominant compared to the other sporadically identified bacterial species (p < in the remaining samples). E. coli was identified alone in 4/20 samples and in association with other bacteria in the other 16/20 peritoneal samples, especially T. pyogenes. The other isolated bacteria species were always associated with T. pyogenes or E. coli, except for Helicoccus ovis and Streptococcus mitis, which were identified alone. Aerobic bacteria were more frequently identified than anaerobic (p < 0.0001); at least one aerobic bacteria (mainly T. pyogenes and E. coli) was cultured in all positive samples (59/59), while anaerobic bacteria (exclusively Clostridium perfringens) were observed in only 4/59 samples and were always associated with aerobic bacteria. All the results of bacteriological culture are summarized in the Scheme 1. Antibodies against at least one of three germs were detected in the majority of blood samples (61/72); only 11/72 were fully negative (p < 0.0001). Antibodies of BoHV4 were the most frequently detected (56/72), compared to those of C. burnetii (15/72)  For BoHV4, the results of ELISA corresponded to those of qPCR in the majority of cases. In other words, in most cows having a negative serology for BoHV4, a negative qPCR was found, and the positive qPCR results corresponded with a positive serology. In rare cases, a positive ELISA result was found in combination with a negative qPCR, or vice versa. For C. burnetii and M. Bovis, a negative result for qPCR and ELISA was observed in the majority of cases. In contrast to BoHV4, several discrepancies between ELISA and qPCR were found; the majority of ELISA positive samples were negative to qPCR, while around half of qPCR positive samples were ELISA negative. The combination of qPCR and ELISA results for the three targeted germs is displayed in Table 1. A positive statistical association was found between the qPCR and ELISA results for BoHV4 and bacteriological culture results. This relation between qPCR and ELISA results and bacteriological culture results could not be confirmed in the case of C. burnetii or M. bovis. The combinations of qPCR and ELISA results and bacteriology outcomes are displayed in Scheme 3. For BoHV4, the results of ELISA corresponded to those of qPCR in the majority of cases. In other words, in most cows having a negative serology for BoHV4, a negative qPCR was found, and the positive qPCR results corresponded with a positive serology. In rare cases, a positive ELISA result was found in combination with a negative qPCR, or vice versa. For C. burnetii and M. Bovis, a negative result for qPCR and ELISA was observed in the majority of cases. In contrast to BoHV4, several discrepancies between ELISA and qPCR were found; the majority of ELISA positive samples were negative to qPCR, while around half of qPCR positive samples were ELISA negative. The combination of qPCR and ELISA results for the three targeted germs is displayed in Table 1. A positive statistical association was found between the qPCR and ELISA results for BoHV4 and bacteriological culture results. This relation between qPCR and ELISA results and bacteriological culture results could not be confirmed in the case of C. burnetii or M. bovis. The combinations of qPCR and ELISA results and bacteriology outcomes are displayed in Scheme 3.

Discussion
This study presents a unique dataset containing a large number of PFP cases observed in the field. Only very few studies have reported aerobic or anaerobic bacteria, BoHV4 and M. bovis in peritoneal liquids and, particularly, in cases of PFP [2][3][4]15]. To our knowledge, the presence of C. burnetii in peritoneal fluids has never been demonstrated.
At least one bacterial species was cultured in more than 80% of the tested samples in this study. This number may even be an underestimation of the true presence, due to the limited sensitivity of bacteriological culture. Also, it is very likely that several cows had been treated with antimicrobials before sampling, modifying the culture results [2,4,16,17].
In a previous publication, anaerobic bacteria have been mainly isolated from the peritoneum during CS [18]. This led to the assumption that bacteria originating from the endogenous vaginal flora and the incised uterus were the main contaminants leading to infectious complications after CS.

Discussion
This study presents a unique dataset containing a large number of PFP cases observed in the field. Only very few studies have reported aerobic or anaerobic bacteria, BoHV4 and M. bovis in peritoneal liquids and, particularly, in cases of PFP [2][3][4]15]. To our knowledge, the presence of C. burnetii in peritoneal fluids has never been demonstrated.
At least one bacterial species was cultured in more than 80% of the tested samples in this study. This number may even be an underestimation of the true presence, due to the limited sensitivity of bacteriological culture. Also, it is very likely that several cows had been treated with antimicrobials before sampling, modifying the culture results [2,4,16,17].
In a previous publication, anaerobic bacteria have been mainly isolated from the peritoneum during CS [18]. This led to the assumption that bacteria originating from the endogenous vaginal flora and the incised uterus were the main contaminants leading to infectious complications after CS. In the current study, in contrast, aerobic bacteria were isolated far more frequently from PFP than anaerobic germs. T. pyogenes and E. coli were predominant, confirming earlier reports on PFP and generalized peritonitis [3].
T. pyogenes and E. coli are ubiquitous in the environment [19,20] and colonize a wide range of tissues and organs [20][21][22]. Hence, it seems logical that exogenous contamination by T. pyogenes and E. coli during CS is a primary cause of infectious complications, including PFP and peritonitis [23][24][25]. Evidently, the risk of complications increases in the case of a massive contamination or immunosuppression [22,26,27]. On the other hand, since healthy cows can have a physiological bacteraemia [21], it is also plausible that PFP is the result of secondary haematogenous infection of a sterile fluid-filled cavity. In conclusion, PFP is in the majority of cases bacterially contaminated, but it remains to be elucidated whether bacteria are primary agents of PFP originating from the environment, the surgeon's hands, the surgical material or the cow's skin or organs [19,20,22], or secondary contaminants of an initially sterile process.
A large number of cows suffering from PFP displayed a positive serology and/or qPCR for one or more of three emerging pathogens, i.e., BoHV4, C. burnetti and M. bovis. These three germs can invade multiple tissues [13] and share the potential to invade white blood cells, allowing them to escape the host's immune response and to pass into a dormant phase [28][29][30]. Stress, parturition and inflammatory processes can reactivate a dormant infection [2,29,31]. Hence, their presence in the PFP fluids may be the result of haematogenous spread via immune cells toward an inflammatory site, and their true implication in the pathogenesis of PFP remains to be elucidated.
The combination of serology and qPCR results allows some interpretation. Animals with an active infection will typically have a positive serology and high amounts of DNA in the infection site [14,32,33]. A positive qPCR in the absence of blood antibodies is indicative of a recent infection; the time between primary infection and detectable antibody levels ranges from 10 days to 4 weeks for the three germs [29,32,34]. A negative qPCR in the presence of antibodies indicates an inactive infection, inhibition of reactivation by a serologic response [29,33,34], or intermittent bacterial replication [35]. There is a long persistence of antibodies in blood circulation after primary infection [2,36].
Of the three studied germs, BoHV4 yielded the most positive results: over 75% of blood samples were seropositive. This is in line with the endemic situation of this virus in Belgian herds, particularly in beef cattle and older cows [14,37]. For comparison, the seroprevalence has been reported to be 67.5% in aborting cows in Wallonia [14]. Nearly 70% of peritoneal fluids were highly positive for qPCR, mostly in association with a strongly positive serology, indicating an active or reactivated infection. A negative qPCR in combination with a (highly) positive ELISA result was observed in a small number of cases, indicating latency of the virus and a serological response. Since a direct relation between BoHV4 detection and specific lesions has never been established [38], the relevance of BoHV4 in the pathogenesis of PFP remains unclear.
Antibody or DNA detection for C. burnetii and M. bovis yielded far fewer positive results than BoHV4. C. burnetii and M. bovis are common in Belgium: 57.8% of tested herds in Wallonia have been reported to have seropositive animals for C. burnetti, and 30% of herds contain animals actively excreting the germ [33]. Between 2012 and 2016, the apparent herd seroprevalence for M. bovis has been estimated to be 23.6% [13]. The combination of a negative serology and a negative qPCR in peritoneal fluid was most frequently found. A positive ELISA in combination with a negative qPCR was detected in a number of cases, indicating an inactive infection or a serological inhibition of reactivation [29,33]. As for M. bovis, this may also be due to the intermittent bacterial replication [35].
A positive association was observed between the presence of BoHV4 (DNA and/or antibodies) and the bacteriological results. Moreover, C. burnetii and M. bovis antibodies and DNA were rarely found alone, as reported elsewhere [36,39], and were almost always associated with BoHV4 antibodies and DNA. It can be postulated that a decrease of immunity induced by a BoHV4 infection may increase the risk of bacterial co-infection [28,38,40,41]. Inversely, the inflammatory condition caused by bacterial infection might also induce BoHV4 reactivation [40,41].
It should be stressed that the presence of germs, their genetic material or their antibodies in PFP cows does not prove a causal mechanism. Their exact role in the pathogenesis of PFP requires further studies. The presence of peritoneal fluids in matched negative control cows could have shed more light on the importance of a positive bacteriology or qPCR, but this was not feasible in the current study setup.

Conclusions
PFP is a frequent pathology in Belgium. Our study clearly demonstrates that PFP can no longer be considered as a sterile process. Our study confirms previous reports of M. bovis in the peritoneal fluid of cows presenting PFP and adds the PFP as new target sites for BoHV4, C. burnetii and other bacterial species. These germs can colonize the PFP through endogenous and exogenous contaminations of CS or via haematogenous spread. Their exact role in the pathogenesis of PFP cannot be concluded from this dataset and requires further studies.