Penicillin and Gentamicin Concentrations in the Uterine Fluid of Non‐Pregnant Mares Following a Single Intrauterine Infusion

ABSTRACT

Despite their widespread clinical use, there is limited pharmacokinetic data for many equine intrauterine antimicrobials. This study aimed to measure the concentration of gentamicin and penicillin in the uterine fluid of mares following infusion of either a standard (PPGent) or long‐acting (LA‐PPGent) compounded formulation. We hypothesized that both formulations would result in therapeutic concentrations, with total concentrations sustained for longer using the long‐acting formulation. Mares were administered 2400 mg of procaine penicillin and 200 mg of gentamicin via a single intrauterine infusion in either a standard (n = 6) or a lyophilized formulation suspended in a slow‐release matrix (n = 6). Intrauterine fluid was collected over a 72‐h period and analyzed for antibiotic concentrations using high‐performance liquid chromatography and ultra‐high‐performance liquid chromatography and tandem mass spectrometry. Mean maximal concentrations were seen at 0.5 h in group PPGent (Penicillin: 10,123.0 ± 4298.0 μg/mL, Gentamicin: 3397.3 ± 1338.5 μg/mL) and exceeded MIC for relevant organisms for 72 h (Penicillin: 2.59 ± 6.34 μg/mL, Gentamicin: 2.14 ± 2.4 μg/mL). Interestingly, maximal concentrations were lower in group LA‐PPG (Penicillin: 2213.8 ± 967.8 μg/mL—p < 0.05, Gentamicin: 1859 ± 2413 μg/mL) and exceeded MIC for a shorter period of time than the unmodified mixture of commonly used FDA‐approved antibiotics.

1. Introduction

Intrauterine antibiotic therapy is widely accepted for the treatment of bacterial endometritis prior to and at the time of breeding in mares (Pascoe 1995; Causey 2006; Dasciano 2011; Katila 2016). While systemic and tissue concentrations have been described for select antimicrobials, intrauterine pharmacokinetics and antimicrobial activity are unknown for many commonly used medications following either local or systemic administration (Caudle et al. 1983; Bennet 1986; Boyd and Allen 1987, 1989; Santschi and Papich 2000). This is particularly relevant in equine reproductive practice as it is believed that the primary source of infectious bacterial disease is luminal intrauterine organisms that gain access to the uterus from ascension through the cervix. Thus, tissue concentrations may not be relevant to a successful treatment outcome.

Our laboratory is focused on the identification of antibiotic therapies that can achieve suitable intrauterine luminal concentration in the equine uterus and also maintain a high level of antimicrobial activity in the face of the adverse environmental influences of mucus, cellular debris, and leucocytic activity in the infected uterus. Ex vivo studies have demonstrated that a combination of penicillin G procaine (referred to as procaine penicillin throughout our paper and colloquially) and gentamicin is highly active in purulent equine uterine fluid and capable of completely eliminating bacterial growth for a minimum of 24 h (McKelvey et al. 2014; Von Dollen et al. 2019). Furthermore, tissue concentrations of penicillin within endometrial tissue have previously been described after intrauterine administration and are considered adequate to inhibit bacterial growth (Bennet 1986).

Thus, the overall objective of the current study was to determine whether procaine penicillin and gentamicin represent a safe and effective treatment option for intrauterine infusion in mares. Our specific primary aim was to document intrauterine concentrations of procaine penicillin and gentamicin after a single intrauterine infusion of either a compounded formulation or a lyophilized long‐acting compounded formulation of these antibiotics. We hypothesized that both drug formulations would result in therapeutic concentrations (at or above the minimal inhibitory concentration (MIC)) in uterine fluid after a single transcervical infusion. We further hypothesized that penicillin and gentamicin concentrations would be maintained longer using the long‐acting formulation than the standard formulation. Although fluid dynamics and clearance of antibiotics in this environment may affect commonly utilized pharmacokinetic data, we additionally characterized the pharmacokinetic parameters for penicillin and gentamicin in uterine fluid after intrauterine administration of two different formulations. Our secondary aim was to determine whether administration of one or both of the selected formulations were associated with inflammation which may compromise uterine health. Based on clinical use of these antibiotics and preliminary work with the commercial matrix in our laboratory, we hypothesized that neither formulation would result in inflammation or fluid accumulation in the uterus.

2. Materials and Methods

This experimental study enrolled 10 healthy pony mares, one Quarter Horse mare, and one Thoroughbred mare. Mares were between 4 and 15 years of age and weighed between 600 and 1200 lbs. All mares were maintained on pasture with supplemental hay and pelleted ration on a university‐owned farm in Central North Carolina. All procedures were carried out in accordance with North Carolina State University's Animal Care and Use Committee's guidelines for the humane treatment of research animals (IACUC # 18‐077‐O). All mares had a complete breeding soundness examination prior to enrollment in this study, including an ultrasound examination of the uterus, uterine culture, uterine cytology, and uterine biopsy. Mares were determined to be cycling and free of any reproductive or systemic disease at enrollment. At the onset of the study, mares were randomly assigned into two different groups to receive either procaine penicillin and gentamicin (PPGent) or a novel, long‐acting formulation of lyophilized procaine penicillin and gentamicin (LA‐PPGent) suspended in a slow‐release matrix.

2.1. Drug Preparation and Administration

All study mares were examined three times weekly to track their reproductive cycles. After enrollment in the study, mares determined to be in diestrus were administered 5 mg of PGF2alpha (Lutalyse, Zoetis, Parsippany, NJ, USA) to induce luteolysis and the return to estrus. On the first documented day of estrus, based on the absence of a corpus luteum, the presence of a large follicle, and the presence of uterine edema, mares were infused with a single dose of the antibiotic combination (denoted as Day 0). To do this, mares were placed in stocks with tails tied laterally, and their perineum was washed with an iodine‐based soap. Six mares (group PPGent) were administered a single transcervical infusion containing 2.4 million IU (2400 mg) of procaine penicillin and therefore 1440 mg total penicillin G (Bactracillin G; Aspen Veterinary Resources Ltd., Liberty, MO, USA) and 200 mg of gentamicin (GentaVed 100, VEDCO Inc., Saint Joseph, MO, USA) in a total of 10 mLs. PPGent was prepared on‐site by mixing available commercial products in a 4:1 combination (8 mLs of procaine penicillin and 2 mLs of gentamicin). Six mares (group LA‐PPGent) were administered a pre‐prepared formulation of the same antibiotics in a biodegradable polymer matrix in a total of 10 mLs. This formulation was prepared by lyophilizing (Labconco FreeZone 2.5‐L, Labconco, Kansas City, MO, USA) a mixture of 2.4 million IU of procaine penicillin (Bactracillin G; Aspen Veterinary Resources Ltd., Liberty, MO, USA) and 200 mg of gentamicin (GentaVed 100, VEDCO Inc., Saint Joseph, MO, USA) and suspending the lyophilized compound in a liquid biodegradable polymer matrix (US Patent 8,187,640; SR Veterinary Technologies LLC, Windsor, CO, USA) under sterile conditions to create a total volume of 10 mLs. The resultant suspension was refrigerated until use at 3°C. Immediately before infusion, the antibiotics were drawn into a single 20 mL syringe and mixed thoroughly by gently inverting the syringe several times. The infusion was drawn into an insemination pipette and passed through the vagina and the external cervical os under manual guidance. Antimicrobials were deposited into the uterine body in all mares.

2.2. Clinical and Ultrasound Evaluation

Clinical evaluation of mares consisted of daily physical and transrectal ultrasound examination using a 7‐10‐MHz linear transducer (Micromaxx machine, FUIJIFILM Sonosite Inc., Bothell, WA). Ultrasound examinations consisted of evaluation of the size of the largest follicle measured in millimeters, degree of uterine edema graded from 0 to 4 (0, no edema; 1, slight edema; 2, moderate edema; 3, heavy edema; and 4, severe or pathologic edema), and the largest amount of intraluminal fluid present measured in cm graded 0 to 4 (0, no fluid; 1, < 1 cm; 2, 1–2 cm; 3, 2–3 cm; 4, > 3 cm). Each mare's cervix was graded on the degree of relaxation from 1 to 3 (1, one finger width; 2, two finger widths; 3, three or greater finger widths).

2.3. Uterine Fluid and Biopsy Collection

Uterine fluid was collected using a modified technique originally described by Foster et al. for the collection of fluid within the bronchi of calves (Foster et al. 2016). Specifically, uterine fluid was collected by the passage of sterile umbilical tape (Jorgensen Laboratories Inc., Loveland, CO, USA) into the uterine body under digital guidance at 0.5, 2, 4, 8, 24, 32, 48, and 72 h following infusion. A double‐guarded instrument with swab removed (Har‐vet double guarded cytology brush, MAI Animal Health, Elmwood, WI, USA) was manually guided into the vagina to the level of the cervix. The inner guard was introduced into the external cervical os, and a pre‐weighed knotted piece of sterile umbilical tape approximately 1 inch in length and ¼ inch in diameter was introduced into the uterine body using flexible biopsy forceps (EndoJaw Large Disposable Biopsy Forceps, Olympus, Breinigsville, PA, USA). The umbilical tape was held within the uterine body for 30 s, retracted into the inner and outer guards to avoid fluid contamination from the vagina, and withdrawn, protected by a gloved hand. The umbilical tape was immediately weighed to obtain a wet weight, placed in a cryovial, and submerged in liquid nitrogen. Cryovials were stored at −80°C until pharmacokinetic analysis.

Endometrial biopsy samples were collected prior to antibiotic infusion and 8, 24, 48, and 72 h after administration. Briefly, a sterile Jackson uterine biopsy instrument was introduced through the cervix under manual guidance (Jorgensen Laboratories Inc., Loveland, CO, USA). Transrectal palpation was then performed to guide the closed forceps into the cranial uterine body at the level of the uterine bifurcation to obtain a biopsy sample.

2.4. Antimicrobial Extraction

Fluids collected from the mares using umbilical tape had to first be extracted to obtain fluid for high‐performance liquid chromatography (HPLC) and ultra‐high‐pressure liquid chromatography with mass spectrometry (UHPLC–MS/MS) analysis. Each strip was treated with 1 mL of distilled water, soaked for 30 s, and vortexed. The liquid supernatant was collected and divided into two 500 μL aliquots for penicillin and gentamicin analysis.

2.5. Penicillin Analysis

Drug concentrations of penicillin from the umbilical tape extracts were analyzed by HPLC using methods modified from previous penicillin studies in our laboratory (Boison et al. 1992). The process was validated using blank uterine fluid from horses that had not received either medication, and the ICH validation guidelines for analytical methods (http://www.ich.org/products/guidelines/quality/quality‐single/article/validation‐of‐analytical‐procedures‐text‐and‐methodology.html) were followed. Penicillin potassium pure analytical reference standard (potency corrected for the weight of potassium) was dissolved in distilled water to prepare a spiking solution for the calibration curve and quality control (QC) samples. The calibration curve consisted of eight fortified samples, plus a zero concentration (blank) sample, across a range of 0.05 μg/mL to 50 μg/mL of penicillin. Retention time for penicillin was approximately 3.5–4 min.

Supernatant extracted from umbilical tapes was injected directly for HPLC analysis. The final concentration of penicillin in the fluid collected was calculated by factoring in the fluid obtained, the dilution factor from extraction, and the recovery. Separation of peaks for umbilical tape fluid extracts was achieved at 40°C with a Zorbax RX‐C8 4.6 × 150 mm (Agilent Technologies, Wilmington, DE, USA) reverse phase column. The system consisted of an Agilent 1100 series quaternary solvent delivery system, an Agilent 1100 series autosampler, an Agilent 1200 series UV detector with ultraviolet absorbance set at 220 nm, and an Agilent OpenLAB software suite for data collection and analysis (all from Agilent Technologies, Wilmington, DE, USA). The mobile phase consisted of 40% acetonitrile and 60% water, with 0.1% trifluoroacetic acid used as a mobile phase modifier. The mobile phase was filtered and degassed prior to use and prepared fresh. The flow rate was 1 mL/min.

2.6. Gentamicin Analysis

Drug concentrations for gentamicin were analyzed using UHPLC–MS/MS after precipitation of uterine fluid proteins with trichloroacetic acid. The UHPLC consisted of an Ultimate 3000 Pump and Column Compartment, and Autosampler (Thermo Scientific, San Jose, CA, USA) coupled with an Orbitrap mass spectrometer (Q Extractive Focus, Thermo Scientific, San Jose, CA, USA). Briefly, 75 μL of uterine fluid was mixed with 15 μL of 30% trichloroacetic acid (TCA) in water to precipitate uterine fluid proteins. The sample was centrifuged for 5 min at 7500 rpm (6000 × g) to sediment the protein pellet. 50 μL of uterine fluid supernatant was pipetted into autosampler vials and diluted 20‐fold with 950 μL of 50% acetonitrile in water containing 0.1% formic acid and d3‐gentamicin as an internal standard at a concentration of 625 ng/mL. The samples were centrifuged at 2400 rpm (2000 × g) prior to analysis. Calibration spikes, quality controls, and blanks were prepared in 0.1% formic acid in water. The analysis was performed by hydrophilic interaction chromatography (HILIC) with a ZIC HILIC column, 150 mm × 2.1 mm, 5 μm particles (Merck KGaA, Darmstadt, Germany). The mobile phases consisted of A: 50 mM ammonium formate, 1% formic acid in water at a pH of 3.25, and B: 1% formic acid in acetonitrile. The column was maintained at a temperature of 40°C, and the autosampler was set to a temperature of 10°C. The analysis used an injection volume of 7.5 μL. The starting solvent composition was 40% aqueous (A) at 0.350 mL/min, which was increased linearly to 85% A in 1.5 min after injection. This solvent composition and flow rate were maintained at 85% A for 5 min prior to re‐equilibration to 40% A. The flow rate during this 4‐min time period was 0.40 mL/min. All gentamicin components (C1, C2/C2a, C1a) and d3‐gentamicin C1 were eluted from the ZIC HILIC column at 4.9 min. Parallel reaction monitoring in the positive electrospray ion mode was used for analyte detection with a spray voltage of 2.0 kV. The precursor ions of the analytes were determined by the instrument software from the molecular formulas. The formulas are gentamicin C1 as C21H43N5O7 or m/z of 478.3235, gentamicin C2/2a as C20H41N5O7 or m/z of 464.3079, gentamicin C1a as C91H39N5O7 or m/z of 450.2922, and d3‐gentamicin C1 as C21H40D3N5O7 or m/z 481.3423. A collision energy of 20 eV (eV) was used to fragment all the gentamicin analytes within the collision cell. Three fragment ions were used to quantify each analyte species, with two of the ions being identical for each gentamicin component. The identical fragment ions for the gentamicin analytes C1, C2/2a, and C1a were at 160.097 and 322.197 m/z. The third fragment ion was m/z 157.134 r gentamicin C1, 143.118 m/z for gentamicin C2/2a, and 129.102 m/z for gentamicin C1a. Five fragment ions were used to quantify the d3‐gentamicin C1 internal standard at 158.140, 161.103, 165.120, 324.120, and 325.215 m/z.

Raw MS data was batch processed in the Xcaliber software (Thermo Scientific, San Jose, CA, USA). Analytes were identified and integrated, calculating an internal standard‐based calibration curve using a weighted (1/X) linear fit. Uterine fluid concentrations of gentamicin in unknown samples were calculated by the Xcalibur software based on the calibration curve, and total gentamicin concentrations were calculated through the sum of all three components (C1, C2/C2a, and C1a). Twelve calibration spikes were prepared, covering the total gentamicin concentration range of 0.1–500 μg/mL. Calibration curves for gentamicin C1 and C2/2a exhibited a correlation coefficient (r ²) of 0.995 and 0.985 for C1a. The limit of quantitation (LOQ) of the analysis was 0.06 μg/mL for gentamicin C1, and 0.02 μg/mL for C2/2a and C1a. The limit of detection LOD was 0.01 μg/mL for gentamicin C1, and 0.005 μg/mL for C2/2a and C1a.

2.7. Pharmacokinetic Analysis

A descriptive pharmacokinetic analysis of total penicillin and gentamicin within uterine fluid from both groups was performed using a statistical moment, non‐compartmental approach in commercial software (Pkanalix, Monolix Suite 2023R1, Lixoft, France). Standard pharmacokinetic parameters were generated as follows: area under the concentration‐time curve from 0 to infinity (AUC_inf), maximum concentration (C_max), mean residence time (MRT) calculated as the ratio of the area under the moment curve and the area under the concentration‐time curve from 0 to infinity, half‐life (T_1/2), and the elimination rate constant (k_el).

2.8. Endometrial Histopathology

Tissue samples were placed into 10% neutral buffered formalin for at least 25 h, embedded in paraffin, and processed routinely for hematoxylin and eosin staining. Histologic architecture of samples collected at 0, 8, 24, 48, and 72 h was evaluated and classified using standard histologic endometrial grading systems (Kenny and Doig 1986; Snider et al. 2011). Biopsy grades were converted to numbers to perform statistical analysis (grade 1, 1; grade IIa, 2; grade IIb, 3; and grade III, 4). At each time point, the number of neutrophils within the superficial submucosa and endometrial epithelium was assessed at 60× and averaged across 10 randomly chosen fields.

2.9. Statistical Analysis

Clinical data parameters (uterine edema score, uterine fluid score, and cervical tone score), mean penicillin and gentamicin concentrations, endometrial biopsy, and endometrial neutrophil scores were assessed between groups and over time using a repeated measures analysis with treatment, time, and their interaction in the model statement (Statistix 10, Statistix Inc., Tallahassee, FL). For all data, a test for normalcy with a Shapiro–Wilk normality test with a significance level set at 0.05 failed. Consequently, data were additionally tested for sphericity using Greenhouse‐Geisser and Mauchly's equations, and corrected p values were reported where appropriate. Bonferroni All‐Pairwise comparison tests were utilized to compare between time points and groups when a statistical difference was detected.

3. Results

3.1. Clinical Ultrasound Findings

The size of the largest follicle on Day 0 of the experiment ranged from 20 to 50 mm. Two mares ovulated on Day two and two on Day three, whereas eight ovulated after the last examination. Clinical and ultrasound evaluation scores of uterine edema, uterine fluid, and cervical tone are summarized in Table 1. The degree of uterine edema did not differ significantly over time or between groups (p: 0.61). Intrauterine fluid scores were greater at 24 h and 48 h compared to baseline, with 1–2 cm of luminal fluid depth on average compared to less than 1 cm (p: 0.0036), but did not differ significantly between groups (p: 0.6483). Uterine fluid had returned to baseline by Day 3. Antibiotic infusion of either formulation was not associated with any complications. Specifically, colic, inappetence, straining, perineal alopecia, or perineal ulceration were not detected in any mare at any time point during this study.

TABLE 1.

Mean (SD) score for clinical evaluation of uterine edema, uterine fluid, and cervical tone in groups PPGent and LA‐PPGent on different days after intrauterine infusion of either a standard or long‐acting formulation of procaine penicillin and gentamicin.

Parameter	Group	Baseline	Day 1	Day 2	Day 3
Uterine edema 1	PPGent	1.83 (1.33)	1.83 (1.17)	1.50 (1.05)	1.67 (1.21)
	LA‐PPGent	1.00 (0.89)	1.67 (0.82)	1.50 (0.84)	1.50 (0.84)
Uterine fluid 2	PPGent	0.50 (0.84)a	1.00 (1.26)b	1.83 (1.72)b	0.83 (0.75)a,b
	LA‐PPGent	0.17 (0.41)a	1.67 (1.37)b	1.00 (1.55)b	0.33 (0.52)a,b
Cervical tone 3	PPGent	2.17 (0.75)	2.5 (0.54)	2.5 (0.55)	2.33 (0.81)
	LA‐PPGent	2.33 (0.82)	2.50 (0.55)	2.50 (0.55)	2.50 (0.55)

Note: Baseline, before treatment; day 1, 24 h after treatment; day 2, 48 h after treatment; day 3, 72 h after treatment.^a,b values within same row with different superscript are significantly different (p < 0.05).

Abbreviations: LA‐PPGent, long‐acting formulation; PPGent, standard formulation; SD, standard deviation.

¹ Uterine edema score: (0, no edema; 1, slight edema; 2, moderate edema; 3, heavy edema; and 4, severe or pathologic edema).

² Uterine fluid score: 0, no fluid; (1, < 1 cm; 2, 1–2 cm; 3, 2–3 cm; 4, > 3 cm).

³ Cervical tone score: (1, one finger width; 2, two finger widths; 3, three or greater finger widths).

3.2. Uterine Fluid Collection

Each strip of umbilical tape was weighed before and after fluid collection. The weight difference indicated the volume collected in the umbilical tape. The collected amount of intrauterine fluid was small and variable between mares and collection periods (Mean ± SD:150.4 μL ± 72.58 μL, Range: 34.4–335.7 μL; Table S1).

3.3. Pharmacokinetics

Non‐compartmental penicillin and gentamicin pharmacokinetics in uterine fluid after a single intrauterine infusion are summarized in Table 2. Although antibiotic fluid concentrations were not measured prior to antibiotic infusion (T0), the treatment history of these animals was known, and none had received any form of antibiotic in the preceding 6 weeks prior to the onset of the trial, and intrauterine fluid presumably contained no penicillin or gentamicin. As with retrieved fluid volume, the amount of antibiotic retrieved was variable. In several cases at sampling times ≥ 24 h, antibiotic was not detected at one or more collection times and then detected again prior to the conclusion of the sampling period (Tables S2 and S3). Mean penicillin and gentamicin concentration over time curves are depicted in Figure 1. Mean luminal penicillin and gentamicin concentrations were highest at initial sampling (0.5 h after infusion) in both groups. Peak (initial) concentrations of both antibiotics were higher in group PPGent than in group LA‐PPGent (Penicillin: 10,123.0 ± 4298.0 μg/mL vs. 2213.8 ± 967.8 μg/mL—p < 0.05, Gentamicin: 3397.3 ± 1338.5 μg/mL vs. 1859 ± 2413 μg/mL—p > 0.05) despite the fact that the same amount (concentration and volume) of each antibiotic was infused. Mean penicillin concentrations were significantly higher in group PPGent compared to LA‐PPGent at 0.5, 2, and 4 h following administration (p < 0.001; Figure 1a). No statistically significant differences were seen in mean concentrations of gentamicin between groups at any time (Figure 1b).

TABLE 2.

Intrauterine fluid non‐compartmental pharmacokinetic analysis parameters [median (interquartile range)] for penicillin and gentamicin following a single intrauterine infusion in regular (PPGent) and long‐acting (LA‐PPG) formulations of penicillin and gentamicin.

	Penicillin		Gentamicin
	PPGent	LA‐PPGent	PPGent	LA‐PPGent
AUCinf	44,529.00	15,354.00	5456.54	4669.07
(mg/L × h)	(19,436.00–50,917.00)	(3606.00–19,494.00)	(3036.99–10,456.24)	(1086.06–8405.23)
Cmax	11,474.00	3489.00	3430.40	2765.90
(mg/L)	(10,909.00–14,564.00)	(2376.00–3975.00)	(2937.15–3894.06)	(359.02–2981.50)
MRT	2.90	3.10	4.27	3.48
(h)	(1.50–4.20)	(2.30–4.40)	(3.12–4.44)	(2.68–5.00)
kel	0.51	0.42	0.06	0.25
(h−1)	(0.12–0.83)	(0.19–0.45)	(0.02–0.08)	(0.12–0.49)
T1/2 (kel)	1.50	1.70	12.55	2.84
(h)	(0.84–5.60)	(1.50–3.60)	(8.52–37.82)	(1.43–5.88)

Abbreviations: AUC, area under the curve; C_max, maximum concentration; k_el, elimination rate constant; MRT, mean residence time; T_1/2, half‐life.

FIGURE 1.

Mean (SD) concentration of penicillin (a) and gentamicin (b) in uterine fluid of non‐pregnant mares after a single intrauterine infusion of a standard (PPGent) or long‐acting compounded formulation (LA‐PPGent).

Mean intrauterine fluid penicillin concentrations remained above the CLSI‐approved breakpoint or MIC of susceptible isolates of Streptococcus equi subspecies zooepidemicus of 0.5 μg/mL at all time points over 72 h in group PPGent (trough Penicillin: 2.59 ± 6.34 μg/mL) and for the first 24 h after administration in group LA‐PPGent, but a high degree of variability was found between mares and collection timepoints (Table S2). Mean intrauterine gentamicin concentrations remained above the CLSI‐approved breakpoint or MIC of susceptible isolates of Escherichia coli of 2 μg/mL at all time points for group PPGent (trough Gentamicin: 2.14 ± 2.40 μg/mL). Mean intrauterine gentamicin concentrations were also above the MIC of Escherichia coli in group LA‐PPGent for 32 h after administration. As gentamicin is a concentration‐dependent antimicrobial, additional analysis was performed to confirm concentrations greater than 8–10 times the MIC of 2 μg/mL for gentamicin. Concentrations of gentamicin in uterine fluid were highest at 0.5 h (mean 1300 × MIC) after infusion and were maintained above the accepted cutoff of 8–10 × MIC for at least 4 h in all mares in both groups (Figure 1b). As for penicillin, a high degree of variability was seen between mares and sampling times (Table S3).

3.4. Histopathology

Endometrial biopsy grade did not significantly differ over time or between groups. No difference was seen between groups in the neutrophil score (p: 0.61); however, a time effect was noted (p: 0.031) in both groups, with a temporary significant increase in the number of neutrophils seen at 8 h compared to baseline or subsequent sample times (Table 3).

TABLE 3.

Mean (SD) endometrial biopsy score and numbers of neutrophils per high power field in both groups on different days after intrauterine infusion of either a standard or long‐acting formulation of procaine penicillin and gentamicin.

Parameter	Group	Baseline	8 h	Day 1	Day 2	Day 3
Biopsy score 1	PPGent	2.17 (0.75)	2.33 (0.52)	2.33 (0.52)	2.40 (0.55)	2.33 (0.52)
	LA‐PPGent	2.33 (0.51)	2.33 (0.52)	2.33 (0.52)	2.50 (0.84)	2.33 (0.52)
Neutrophil score	PPGent	1.98 (4.71)a	17.27 (17.86)b	8.48 (8.90)a,b	4.40 (2.18)a,b	11.97 (11.67)a,b
	LA‐PPGent	0.12 (0.29)a	10.25 (8.21)b	8.52 (8.29)a,b	8.00 (7.80)a,b	8.53 (6.87)a,b

Note: Baseline, before treatment; 8 h; 8 h after treatment; day 1, 24 h after treatment; day 2, 48 h after treatment; day 3, 72 h after treatment.^a,b values within same row with different superscript are significantly different (p < 0.05).

Abbreviations: LA‐PPGent, long‐acting formulation; PPGent, standard formulation; SD, standard deviation.

¹ Biopsy score: (grade 1, 1; grade IIa, 2; grade IIb, 3; and grade III, 4).

4. Discussion

In combination with other published work from our laboratory and others, this study supports the use of procaine penicillin and gentamicin as an intrauterine therapeutic (Bennet 1986; McKelvey et al. 2014; Von Dollen et al. 2019). High concentrations of these agents were achieved in the uterine fluid, and antibiotic concentrations in this environment were maintained above susceptible breakpoints for common gram‐positive and gram‐negative microbes for an acceptable duration of time for both drugs. In fact, mean antibiotic concentrations of penicillin and gentamicin were above published MIC for relevant organisms for up to 72 h after infusion of FDA‐approved commercially available procaine penicillin and gentamicin formulations as a 4–1 premixed solution in estrous mares (group PPGent). Individual variability in antibiotic concentrations between animals and collection periods may be explained by inherent variations in uterine clearance as well as the small and variable volume of fluid obtained with this collection method (Tables S1–S3). Notwithstanding these limitations, penicillin concentrations were maintained above MIC for a minimum of 8 h in all animals (724.26 ± 1285.86 μg/mL). Eight of twelve mares had concentrations exceeding MIC for relevant organisms at 24 h (84.42 ± 297.96 μg/mL), while two additional animals had no detectable penicillin at 24 h, but intrauterine penicillin concentrations exceeding MIC at 36 or 72 h, suggesting that a representative sample may not have been obtained at some timepoints. Gentamicin concentrations reached a mean of 1300 × MIC for relevant organisms after intrauterine infusion, vastly exceeding the recommendations of peak concentrations above 8–10 × MIC for concentration‐dependent antibiotics. Combined with in vitro work, which demonstrated that this combination of antimicrobials prevents the growth of both E. coli and S. zooepidemicus for at least 24 h in equine uterine fluid (McKelvey et al. 2014; Von Dollen et al. 2019), this study supports the commonly utilized once daily treatment interval.

Minor inflammatory changes were noted during the course of this study, which returned to baseline prior to the end of testing at 72 h. Although no untreated catheterized mares were included for comparison, these changes did not rise to the level of clinical relevance and are consistent with minor irritation associated with repeated transcervical manipulation for any purpose (Blanchard et al. 2003). A mild increase in neutrophil score was reported at 8 h, which diminished in subsequent biopsy specimens and did not affect the overall uterine biopsy score at any time point. Likewise, a mild increase in uterine fluid, but not uterine edema was noted on ultrasound exam at 24 and 48 h compared to baseline. Safety of this combination of antibiotics is further supported by extensive empiric clinical use in the field and previously published work in healthy pregnant mares from the same herd (Beachler et al. 2021).

Finally, a consideration that must be weighed before the clinical application of this protocol is the nature of the compounded antibiotic product used. The American Veterinary Medical Association (AVMA), American Association of Equine Practitioners (AAEP), and the Food and Drug Administration (FDA) have discouraged the compounding of antibiotics (and other medications) from bulk chemical substances, as described in FDA GFI #256. However, The Federal Food, Drug, and Cosmetic Act (FD&C Act) permits compounding of animal drugs when the source of the active ingredient is a finished FDA‐approved drug and not a bulk drug substance. A “bulk drug substance” is a substance used to make a drug that becomes an active ingredient in the finished dosage form of the drug. Although any time a veterinarian or pharmacy mixes two substances together, its result is a “compounded product”, this form of compounding is generally permitted and the resulting products still fall within FDA guidelines. In this study, we used the FDA‐approved sources of penicillin and gentamicin for both our formulations. Group PPGent utilized the simple 4–1 combination of procaine penicillin and gentamicin available to practitioners, whereas for group LA‐PPGent, these same FDA‐approved products were lyophilized and resuspended in polymers designed to extend release and delay clearance of the active compounds were added in group LA‐PPGent. In contrast to our initial hypothesis, the addition of a complex polymer to extend bioactivity was ineffective at enhancing or prolonging antibiotic concentration in uterine fluid.

4.1. Conclusions

In summary, this study is the first to demonstrate that a single dose of combined procaine penicillin and gentamicin results in sufficient intrauterine antibiotic concentrations to inhibit the growth of intrauterine Streptococcus equi subsp. zooepidemicus and Escherichia coli in mares and further demonstrates that the use of a compounded formulation adding polymers to prolong the release of antibiotics in the uterus is not necessary or effective. In combination with other studies from our laboratory, it lays the groundwork for clinical application and larger retrospective studies to provide therapeutic guidance for the use of procaine penicillin and gentamicin in the treatment of intrauterine infections.

Author Contributions

Scott C. Bailey and Theresa M. Beachler conceived and designed the study. Theresa M. Beachler, Scott C. Bailey, and Kate Withowski executed the study and performed sample collection. Mark G. Papich, Larry W. Wulf, and Michael Yaeger performed sample analysis. Jonathan P. Mochel and Debosmita Kundu performed pharmacokinetic modeling. All authors were involved in drafting and revising the manuscript.

Ethics Statement

This study was approved by the Institutional Animal Care and Use Committee of North Carolina State University (Protocol # 18‐077‐O).

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Data S1.

Data Availability Statement

All relevant data is presented in this manuscript.