Randomized field trial comparing the efficacy of florfenicol and oxytetracycline in a natural outbreak of calf pneumonia using lung reaeration as a cure criterion

Abstract Background Respiratory infections are the main indication for antimicrobial use in calves. Optimal treatment duration currently is unknown, but shorter duration would likely decrease selection for antimicrobial resistance. Hypothesis/Objectives Determine differences in cure rate and healing time between animals treated with florfenicol and oxytetracycline in a natural outbreak of respiratory disease using reaeration observed on thoracic ultrasound examination as healing criterion. Animals Commercial farm housing 130, 3 to 9 month old Belgian blue beef calves. Methods Randomized clinical trial during an outbreak of respiratory disease. Metaphylactic treatment was initiated, randomly treating animals with either florfenicol or oxytetracycline. Ultrasonographic follow‐up was done the first day and every other day for a 14‐day period. At the individual animal level, treatment was discontinued when reaeration of the lungs occurred. Differences in cure rate and healing time were determined. Results Of the 130 animals studied, 67.7% developed a lung consolidation ≥0.5 cm. The mean ultrasonographic healing time was 2.5 days in the florfenicol group compared to 3.1 days in the oxytetracycline group (P = .04). After single treatment, 80.6% and 60.3% had no consolidations in the florfenicol and oxytetracycline groups, respectively (P = .01). A Mycoplasma bovis strain was genetically and phenotypically determined to be susceptible to both antimicrobials. Conclusions and Clinical Importance Ultrasonographic lung reaeration shows potential as a cure criterion to rationalize antimicrobial use for outbreaks of pneumonia. In our study, florfenicol resulted in a faster cure and higher reduction in antimicrobial usage than did oxytetracycline.


| INTRODUCTION
Respiratory tract infections are a major cause of morbidity, mortality, and antimicrobial use in all cattle industries worldwide, including dairy, beef, and veal farms. 1,2 Mixed infections are often present, consisting of viruses, M. bovis and Pasteurellaceae species. 3 Increasing scientific evidence and consumer concerns regarding antimicrobial resistance selection warrant an urgent reduction in antimicrobial use without hampering animal welfare. 4,5 Metaphylactic antimicrobial group treatments frequently are used to control bovine respiratory disease (BRD) outbreaks. 6,7 This approach is increasingly criticized, especially because many outbreaks include viruses and it is unclear to what extent bacterial superinfection occurs. [8][9][10] A meta-analysis showed mainly short-term effects of metaphylaxis and stated that the number needed to treat is too high for this approach to remain acceptable. 8 However, metaphylaxis likely will remain essential to control some highly contagious bacterial pathogens such as M. bovis, which is involved in approximately one third of respiratory disease outbreaks in dairy and beef calves and in almost in every veal calf or feedlot outbreak. 9,11,12 Thoracic ultrasound (TUS) potentially can reliably identify animals with pneumonia (either clinical or subclinical) and differentiate them from those with upper respiratory tract infections. [13][14][15] In addition, individualization of treatment length based on TUS may substantially decrease antimicrobial use, even when metaphylaxis is used. In practice, a 5 to 10 day antimicrobial treatment length for pneumonia often is used, but scientific evidence is limited and criteria for success are poorly definded. [16][17][18] It is unknown whether required treatment duration would differ among antimicrobial classes. Both European Medicines Agency (EMA) and World Health Organization (WHO) recommend more restricted use of macrolides, fluoroquinolones and cephalosporins, whereas florfenicol (FF) and oxytetracycline (OTC) are most frequently mentioned as first choice empirical treatment for BRD in formularies from different European countries. 5,[19][20][21] Historically, criteria for cure have been based on the disappearance of clinical signs linked to respiratory disease. 10 On TUS, pneumonia is characterized by the presence of ≥1 lung consolidations, and in studies of humans reaeration of a previously consolidated area is considered a sign of healing. 22,23 It is unknown whether lung reaeration can be used as a criterion to discontinue antimicrobial treatment in cattle.
Therefore, our objective was to assess the possibility of using lung reaeration as an individual animal cure criterion to discontinue antimicrobial treatment. Second, differences in cure rate and required treatment duration between FF and OTC were determined. As a third objective, susceptibility of the involved M. bovis strains was determined both by microdilution and genome sequencing.  24 To maximize diagnostic accuracy, scanning was not done at the usual qTUS speed of 1 to 2 minutes per animal. Briefly, the scanning technique consisted of 1 single movement on each side of the thorax starting from the tip of the diaphragmatic lobes to the tip of the cranial lung lobes. In larger calves, 2 movements on each side often are needed to visualize the entire lung. After being positioned parallel with the ribs, the probe is moved cranially from intercostal space to intercostal space, without losing contact with the skin. This position maximizes the lung area visualized and minimizes acoustic shadowing of the ribs. The landmark to initiate the scan for the right side is an image containing the lung dorsally and the liver ventrally (8th-9th intercostal space). On the left side, (8th-9th intercostal space) it is an image with the lung dorsally, the spleen in the middle and the rumen ventrally. The second landmark is an image taken at the level of the heart (3rd-5th intercostal space) consisting of the lung dorsally and the heart ventrally in a 50/50 distribution. This landmark is used to reorient the probe. Next, the probe is advanced cranially under the triceps muscle in this position. In this manner, scanning of the cranial lung lobes (mainly the right cranial lung lobe, but visualization of this lobe from the left side After the 14-day trial period, the clinical course of the animals was followed as part of the herd health services provided to the farm, which did not include systematic ultrasonography at the time of the study. The farmer was asked to report signs of decreased growth or carcass weights if divergent from expected results. were not informed to which test group an animal belonged. No externally visible marks were present on the animals to distinguish treatment groups. Once allocated to a treatment group, animals were treated with the same antimicrobial as the previous treatment at the same dose for as long as the lung consolidation was present. Both drugs were administered IM q48h, according to manufacturer recommendations. Body weight was estimated using chest circumference measurements done at first visit and rounded up to the next 10 kg. A consolidation depth ≥0.5 cm was used as a threshold to initiate antibiotic treatment.

| Treatment protocol
During the trial period, animals were scanned every other day, only retreating calves with lung consolidations with a depth of ≥0.5 cm.
Reaeration was defined as the reappearance of reverberation artifacts (A-lines) in previously consolidated lung tissue. When complete reaeration of the lung occurred, treatment was discontinued.

| Sampling and laboratory diagnosis
Before treatment was initiated, a small-volume (30 mL saline) nonendoscopic bronchoalveolar lavage (nBAL) was performed in 5 standing, unsedated calves by a veterinarian, using a sterilized catheter as previously described. 29 Sampled calves were conveniently selected, only assuring they originated from a different pen and showed lung consolidations on TUS. Samples were taken before treatment was initiated. Bacterial cultures for M. bovis were done on all 5 samples using a selective indicative agar (Pleuropneumonia-like organism agar with Tween-80 for lipase activity), as described elsewhere. 30,31 Multiplex real time-PCR (qPCR) for various BRD pathogens was performed on a pooled sample from the 5 nBAL's. 32 Definitive identification of M. bovis was performed using MALDI-TOF MS (Brüker Daltonik GmbH, Bremen, Germany), as described by previously and qPCR. 32 Antimicrobial susceptibility of M. bovis for both FF and OTC was determined using broth microdilution, determining minimum inhibitory concentration (MIC) using previously described techniques. 33 Because of the lack of clinical breakpoints for M. bovis, MIC values were interpreted by determination of the epidemiological cutoff (ECOFF) using the visual method as described previously. 33 The ECOFF values for FF and OTC were reported as >16 and >8, respectively. 33 Strain typing of M. bovis and detection of possible genomic antimicrobial resistance was done by long-read nanopore sequencing using techniques previously described. 34

| Statistical analysis
All data was entered in a spreadsheet (Excel, Microsoft, Inc) and transferred to both SPSS statistics 27 and SAS 9.4 (SAS Institute, Inc, Cary, North Carolina). Graphs were made using GraphPad Prism (version 9.1.1 for Mac, GraphPad Software, San Diego, California, www. graphpad.com). The individual calf was defined as an experimental unit. Primary outcomes of interest were cure rate and healing time.
Cure was defined as showing complete reaeration (reverberation artifacts) of a previously consolidated lung. Similarly, cure rate was defined as the proportion of calves that showed lung reaeration of a previously consolidated lung area. Healing time was defined as the number of days between start of the treatment and appearance of complete lung reaeration, as a criterion for lung healing at which antimicrobial treatment was stopped. Healing time represented the minimum required treatment duration to obtain complete lung reaeration.
In this trial, all animals were considered at risk for development of pneumonia. Therefore, for animals that did not show lung consolidations at the first examination (day 1), the healing duration was set at 2 days because the dosing regimen ensured an effective antimicrobial blood concentration for 2 days after metaphylactic treatment was initiated. Because the use of metaphylaxis also relies on the prevention of developing pneumonia, these animals were counted as part of the total number of healthy animals after treatment. All analyses were done for the entire group (animals with and without lung consolidations on the first day of the trial) and separately for the animals with lung consolidations on day 1. Considering a noninferiority limit of 10%, the available sample size allowed for detection of a 10% difference in cure rate between the 2 groups with 80% power and alpha <.05. Using an expected SD in healing time of 4 days allowed for detection of differences in healing time of 2 days between FF and OTC.
To determine the difference in cure rate, multivariable logistic regression was used with cure as binary outcome variable. The overall test for treatment effect was based on the Chi-squared test (Wald test) and treatments were compared pairwise using Tukey's multiple comparison technique. Each model contained treatment group, time, and interaction of treatment group with time. Pen was forced as a random factor in each model to account for clustering of calves in a pen. In a multivariable analysis, in addition to treatment group sex and clinical signs (temperature, breathing frequency, cough, and apathy) also were evaluated for potential confounding effects on cure.
To bring healing time into account, a Cox proportional hazards model was constructed with cure (0/1) as the outcome variable and treatment group as predictor in a second approach. The time variable was defined as the number of days between initiation of the treatment and discontinuation of treatment when lung reaeration was seen. Right censoring was done at day 14 when the study was ended.
The same predictors as mentioned for logistic regression were included. A linear mixed model was used to determine the difference in required treatment duration until reaeration (i.e., healing time) between FF and OTC. The outcome variable was treatment duration, based on therapeutic blood concentrations (2 days at the long-acting dose used in the study) and the predictor was treatment group. Again, pen was forced as a random factor into the model. Finally, the number of animal daily dosages (ADD) was calculated for each treatment group by adding the number of days that each animal had received antimicrobials, taking the long-acting effect (2 days) of FF and OTC into account. One-sample t testing was used to compare ultrasoundguided treatment to a hypothetical 7-day metaphylaxis. A similar linear mixed model approach as mentioned above was used to determine the difference in ADD between the FF and OTC group. In all models, significance was set at P < .05.

| Animals and outbreak characteristics
To avoid age bias, animals <3 months of age were excluded from the   The genome of 1 of the obtained isolates was screened for previously described mutations in M. bovis involved in antimicrobial resistance as described elsewhere. 35 No mutations directly associated with FF or OTC resistance were detected. However, known mutations for other antimicrobials were observed. A G748A mutation in both 23S rRNA alleles, as well as a Gln93His mutation in the rplv gene were found. In previous studies, G748A mutations have been observed in isolates resistant to tylosin and tilmicosin, whereas the Gln93His mutation has been linked to resistance against macrolides (gamithromycin and tylosin) when combined with a mutation in domain 5 of 1 or both 23S rRNA gene alleles on position A2058. [36][37][38][39] In this case, no mutations on position A2058 were found. Additionally,

| DISCUSSION
Our main objective was to determine whether stopping antimicrobial treatment after complete reaeration of a previously consolidated lung area is an option to achieve adequate cure with less antimicrobial use. After lung reaeration, no animals relapsed in the 14-day follow-up period and in the months after, the farmer did not report decreased carcass weights compared to other groups of animals. However, because ultrasonographic follow-up was not performed after day 14, a relapse with subclinical pneumonia cannot be excluded. Because of the bovine lung anatomy, with few collateral bronchi, a general assumption is that inflammation of a main bronchus more rapidly results in a consolidation of the complete lung lobe in cattle, as compared to species with many collateral bronchi, such as horses. 40,41 In this case, metaphylactic treatment was started in the early stages of the outbreak and a 0.5 cm consolidation depth was used as cutoff for treatment. In these animals, only lobular pneumonia was present, which may explain the rapid reaeration after treatment. In the 12 calves with consolidation of the complete lung lobe (≥3 cm), however, complete reaeration occurred in all but 1 calf. Hence, frequentlyencountered lung consolidations in cattle in the slaughterhouse more likely represent insufficient disease detection (including subclinical presentations) and ineffective treatment, rather than an inevitable consequence of pneumonia. 42 Second, our study aimed to determine the difference in cure rate and healing time between 2 antimicrobials commonly used for pneumonia in calves, FF and OTC. We chose these 2 antimicrobials because of their position as first and second choice in formularies of different EU countries. 21 In M. bovis from Belgian herds, acquired antimicrobial resistance against macrolides is frequent, whereas it is rare for FF and OTC, similar to most reports worldwide. 33 and reliability of the ultrasonographic method, the possibility to train people (potentially also lay staff), and on the economic impact of pneumonia that is not cured. The pressure to decrease antimicrobial use depends on local legislation, but can be a very important incentive to decrease antimicrobial use. The ultrasonographic technique used in our study requires 1 to 2 minutes per animal and is easily learned by novices. 24 After a standard treatment of 5 to 7 days, 10.2% and 5.6% of the animals still were not cured in the FF and OTC groups, respectively. This outcome means that, when a standard 7-day treatment is applied, these animals would still lead to economic losses as previously documented. 2,[43][44][45] Third, in our study, FF resulted in more rapid cure than OTC, In our study, it was surprising that there was a difference in cure rate and healing time between 2 bacteriostatic antimicrobials.
Although bacteriostatic and bactericidal data may provide valuable information on the potential action of antibacterial agents in vitro, it is necessary to combine this information with pharmacokinetic and pharmacodynamic information to provide more meaningful prediction of efficacy in vivo. One reason for the observed difference in cure rate might be the lipophilic nature and wide tissue distribution of FF. 48 Immediately after IM administration of FF, high initial blood concentrations are reached, leading to a rapid initial response. 49 Oxytetracycline has a rather short-term effect and lower lipid solubility (when compared to doxycycline), leading to a more restricted tissue distribution and faster elimination rate. 48 Only 1 study using TUS previously evaluated antimicrobial treatment (tulathromycin) for BRD. 50 Compared to that study, our cure rates using both FF and OTC were very high. 50 Because the pathogens involved and their antimicrobial susceptibility were not determined in the previous study, it is hard to determine whether this difference in cure rate is a consequence of differences between antimicrobials, antimicrobial susceptibility of the involved pathogens or treatment initiation at a lower consolidation depth cutoff in our study.
Regardless of its feasibility under field conditions, the ultrasono- Therefore, we believe chances of encountering multiple strains of M. bovis during this outbreak to be rather limited. Finally, dosage of the antimicrobials was based on chest circumference measured at first visit and rounded to the next 10 kg. Animals were given the same dose throughout the entire trial, not taking daily growth into account.
Calculating required dosage using exact body weights may be beneficial in further decreasing required treatment duration, but is difficult to achieve in every farm.

| CONCLUSIONS AND CLINICAL IMPORTANCE
Lung reaeration on ultrasonography proved to be a promising criterion to discontinue antimicrobial treatment in an outbreak of respiratory disease in beef calves. Personalizing treatment duration by using repeated TUS holds potential to both rationalize antimicrobial use and limit production losses. In our study, FF resulted in a more rapid cure rate than did OTC in the early phase of a natural outbreak of M. bovis, potentially complicated by Pasteurellaceae infection.