Antibiograms of field and hospital acquired equine neonatal bacterial fluid cultures in the Midwestern United States: 149 samples (2007‐2018)

Abstract Background Contemporary data reflecting local pathogens and their antibiograms is necessary to select empirical antimicrobial therapy for equine neonates. Hypothesis/Objectives Describe bacterial isolates associated with equine neonatal infection and their antibiograms in the Midwestern United States. An increase in gram‐positive infection and antibiotic resistance compared to previous literature was expected. Animals Data from 149 fluid samples from 133 foals <30 days of age submitted for bacterial culture between January 2007 and December 2018. Methods A retrospective evaluation of equine neonatal fluid cultures was performed. Fluid submission type, bacterial culture and antibiogram, empirical antibiotic treatment, and foal outcome was included. Isolate susceptibility to individual antimicrobials and combination protocols relevant to equine practice were recorded. The effect of recorded variables on foal survival was evaluated using Fisher's exact or chi‐squared tests. Results Ninety bacterial isolates (78 aerobes and 12 anaerobes) were identified and gram‐positive organisms predominated (n = 50/90, 56%). Greater than 70% of aerobic isolates were susceptible to ampicillin, ceftiofur, chloramphenicol, trimethoprim/sulfamethoxazole, and all penicillin/aminopenicillin and aminoglycoside combinations. Seventy‐seven (n = 81/105) percent of foals survived. Survival was associated with a negative fluid culture and was not associated with empirical antimicrobial choice. Conclusions and Clinical Importance Gram positive and anaerobic isolates associated with equine neonatal fluid cultures exceed that of previous reports. Historical empirical antimicrobial choices for equine neonatal infection in the Midwestern United States are supported by local antibiogram results.


| INTRODUCTION
Equine neonatal sepsis is the combination of demonstrable infection and clinical evidence of systemic inflammation which includes at least 3 of the following: elevated rectal temperature, tachycardia, tachypnea, abnormal leukocyte count (leukocytosis, leukopenia, or band neutrophilia), venous hyperlactatemia, or venous hypoglycemia. 1 Despite advancements in equine critical care, neonatal sepsis remains a frustrating and important cause of mortality. [2][3][4][5][6][7] Survival from sepsis is associated with prompt administration of an appropriate antimicrobial 8,9 ; however, bacterial culture and sensitivity is time consuming and lacks both sensitivity and specificity. 6,10-12 As a result antimicrobials are often empirically chosen, and this is complicated by temporal and regional shifts in relevant bacterial pathogens and their antibiograms. Prescribing veterinarians require a clear understanding of these shifts to make rational antimicrobial choices for foals in their care, but regional specific data are largely unavailable. The objective of this study was to document the bacterial culture and susceptibility results In vitro susceptibility to antimicrobials was recorded for aerobic isolates using broth microdilution method. Briefly, 10 μL of 0.5 McFarland suspension of the bacteria in sterile water was inoculated to 10 mL of cation-adjusted Mueller-Hinton broth. Fifty microliters of this bacterial suspension were added to each well of an equine specific antimicrobial sensitivity testing plate (Equin1F Sensititre, ThermoFisher, Waltham, Massachusetts) and the plate was incubated at 35 C for 16 to 18 hours.
After incubation the minimum inhibitory concentration (MIC) for each antimicrobial was manually read and recorded. Antimicrobials were included for analysis if they were considered practical for equine neonatal use and susceptibility data was reported for at least 50% of all bacterial isolates obtained. Amikacin, ampicillin, cefazolin, ceftiofur, chloramphenicol, enrofloxacin, gentamicin, penicillin, tetracycline, and trimethoprim/sulfamethoxazole were included. Bacterial susceptibility was also reported for antimicrobial combinations common to equine practice and included amikacin/ampicillin, amikacin/penicillin, gentamicin/ampicillin, and gentamicin/penicillin. Susceptibility data was reported in ordinance with Clinical and Laboratory Standards Institute (CLSI) established breakpoints. Multiple drug resistant (MDR) isolates were defined by resistance to 3 or more evaluated antimicrobials. 11 Case data including age, breed, and sex were recorded for all foals, and antimicrobial therapy and survival data were reported for all hospitalized foals. Empirical antimicrobial therapy was defined as the first antimicrobial administered to the foal. This therapy was categorized as "appropriate" if the antimicrobial demonstrated in vitro efficacy against organism(s) isolated from the foal's first fluid culture submission(s). Empirical antimicrobial efficacy was only evaluated if: the initial fluid sample yielded positive growth, an antibiogram was available for the recovered isolate(s) and the specific empirical antimicrobial used was recorded. If serial fluid cultures were submitted, changes in the isolated bacterial agents and their antibiograms were reported but not used to define empirical antimicrobial efficacy. Survival was defined as survival to hospital discharge and was not considered for field cases. Clinical and biochemical case data required to identify systemic inflammation was not evaluated, and enrolled foals with positive fluid cultures could not be defined as septic.
Foal age was tested for normality using a Shapiro test, and year of submission was assessed for uniform distribution. Categorical data included submission type (hospital or field), fluid type, bacterial isolate, antimicrobial sensitivity, empirical antimicrobial efficacy, and survival.
Relationships between these categorical variables were evaluated using a Fisher's exact (2 variables) or chi-squared (>2 variables) tests.

| RESULTS
One hundred forty-nine fluid cultures were performed on 133 foals <30 days of age. This included 124 hospital acquired fluid cultures (n = 108 foals) and 25 field acquired fluid cultures (n = 25 foals).
Seventy-six males and 49 females were enrolled and sex was not recorded in 8 foals. Breeds were representative of the local population and included quarter horse/paint, other/mixed, Arabian, warmblood, standardbred, and thoroughbred most commonly.
Average age at the time of fluid culture submission was 6.6 days (range, 0-29), and age at submission differed between blood and synovial fluid samples (4.6 ± 3 and 10.8 ± 9 days respectively, P = .02).  Table 1 outlines the percent, number, and fluid type for each genus isolated.
Ten foals had multiple fluid cultures performed including blood and synovial fluid cultures (n = 6), multiple synovial fluid cultures (n = 2) repeat blood cultures (n = 1) and blood and CSF cultures (n = 1). For no foal were all isolates identical in repeat samples; for 6 foals the same bacteria were isolated from multiple cultures, but in each of these 6 foals disparate isolates were also recovered.

| Survival
Gram-negative isolates were more common in synovial samples, and gram-positive isolates were more common in blood samples. How-  for field practitioners, although this was not quantified. It is, however, also possible that there are pathophysiologic differences in infection based on gram status in foals as is seen in human medicine 23 and has been suggested in foals. 16  Greater than 70% of recovered isolates were susceptible to ampicillin, ceftiofur, chloramphenicol, trimethoprim sulfamethoxazole, and all antimicrobial combination protocols. There were no significant differences in susceptibility across combination protocols. Chloramphenicol was the most effective individual antimicrobial, with 92% of isolated bacteria demonstrating in vitro sensitivity. However, its use in equine neonates faces pharmacologic and practical limitations. In adult horses repeat oral dosing of chloramphenicol fails to achieve CLSI MIC targets for >50% of the conventional dosing interval. 25 Limited pharmacokinetic data is available on oral administration of chloramphenicol in foals, but the data available suggest mean peak plasma concentrations are highly variable among individuals and unlikely to consistently reach CLSI breakpoints. 26 This could be related to reduced bioavailability associated with gastrointestinal dysfunction in sick foals, 27 though this finding is inconsistent. 24 Nonetheless, without supportive pharmacokinetic evidence, concerns regarding achievable plasma concentrations can dissuade the use of chloramphenicol in foals despite in vitro susceptibility support. Additionally, the small but serious risk of aplastic anemia in humans exposed to chloramphenicol during administration must be considered before prescription. Oral dosing is challenging in foals and amplifies this risk.
In this study, 85% of isolates were susceptible to ceftiofur at the CLSI MIC breakpoint of ≤2 μg/mL. This exceeds isolate susceptibility in geographically distinct studies. 12,14,17,18  Seventy-seven percent of foals survived to hospital discharge and survival was associated with a negative fluid culture as reported. 2,4,10,12,16,18 In contrast to previous reports, survival was not associated with polymicrobial infection 16 or in vitro susceptibility of the bacterial isolate to empirical antimicrobial choices. 9 Inappropriate empirical antimicrobial therapy was identified in 6 out of 24 foals; however, 4/6 of these foals received extra-label ceftiofur sodium at 5 to 10 mg/kg IV every 6 to 12 hours, and in vivo ceftiofur concentrations likely exceeded MIC breakpoints. The association between in vitro efficacy of an initial antimicrobial choice and survival in septic animals is inconsistent. A 1.5-fold increased risk of death is reported when septic foals were treated with empirical antimicrobials that were ultimately deemed inappropriate. 9 However, a recent study in septic dogs found no association between survival and empirical antimicrobial treatment. 32 Other animal or treatment factors, and particularly the differences between in vivo and in vitro antimicrobial susceptibility likely contribute to these inconsistencies. Importantly, survival analysis was not the primary aim of this study, and the association between survival and empirical antimicrobial choice was only evaluated in 25 hospitalized foals. Notably, euthanasia due to financial constraints was not identified and undoubtedly influenced outcomes, particularly because this study period included an economic recession.
This study faces several limitations related to both sample size, retrospective data collection, and inclusion of both hospital and field-based submissions. Incomplete and small data sets prevented some useful observations, such as comparing culture and antibiogram results across the study period, or between hospital and field submissions, which are reported. 7,17,33 Similarly, clinical and biochemical data were not reported for included cases, and without this data sepsis could not be defined in our study population. 1 Differences between hospital and field derived fluid samples also influenced results. Field submissions were collected from older foals and were more likely to be synovial compared to hospital acquired samples. Presumably this is because field cases were more advanced in their disease process and more commonly presented with signs of infectious orthopedic disease compared to systemic infection.
Though these differences challenge direct comparisons between the 2 populations, they also highlight missed opportunities for field practitioners to potentially identify and treat bacteremia in younger sick foals.
Many field practitioners are faced with financial and practical barriers to referral, and early blood culture of sick foals can support informed antimicrobial choices that expand the regional understanding of bacteremia in the nonhospitalized, sick foals. are also reported clinically 24 and were therefore included for analysis.

ACKNOWLEDGMENT
No funding was received for this study. A poster abstract of this work was presented at the 2021 International Veterinary Emergency and Critical Care Conference in Nashville, TN.