Escherichia coli isolated from canine pyometra show decreased susceptibility to amoxicillin–clavulanic acid in dogs
Authors: BVetMed Emailzfautz@gmail.com, DVM, DACVS, and DVM, DACVIM
Abstract
Objective
To describe bacterial culture results and antimicrobial susceptibility patterns for dogs undergoing surgical treatment for pyometra.
Methods
This retrospective study included dogs that underwent ovariohysterectomy at referral hospitals and had a positive bacterial culture of uterine fluid or histopathologic changes consistent with pyometra between January 2020 and December 2024. Data collected from medical records included clinical signs on presentation, hematological and biochemical abnormalities, aerobic and anaerobic bacterial isolates, antimicrobial susceptibility results, and peri- and postoperative antimicrobial administration. Prevalence of antimicrobial susceptibility was compared across select geographic regions and between Clinical and Laboratory Standards Institute (CLSI) susceptibility interpretation guideline editions.
Results
278 dogs were included. Neutrophilia was identified in 74.2% of dogs. Only 1 bacterium was isolated in 88.8% of dogs, while 10.4% of cases were polymicrobial. Escherichia coli was the most common isolate (73.1%). Amoxicillin–clavulanic acid and enrofloxacin were prescribed postoperatively to 77.6% and 51.1% of dogs, respectively. Using CLSI VET01S edition 4, E coli susceptibility to amoxicillin–clavulanic acid and enrofloxacin was present in 30.3% and 96.9% of isolates, respectively. Reclassification using CLSI VET01S edition 7 resulted in lower susceptibility classifications in 9 of 10 antimicrobials evaluated, including reductions of 46.9% for amoxicillin–clavulanic acid and 3.1% for enrofloxacin.
Conclusions
Amoxicillin–clavulanic acid was frequently prescribed despite limited susceptibility among E coli isolates. Updated CLSI breakpoints reclassified susceptibility for most antimicrobials despite unchanged MICs.
Clinical Relevance
Regional antibiograms may show decreased susceptibility compared to historical recommendations. Judicious empiric antimicrobial selection should be guided by antimicrobial stewardship principles rather than routine reliance on any single antimicrobial.
Keywords: amoxicillin–clavulanic acid; antimicrobial resistance; antimicrobial susceptibility testing; Escherichia coli; pyometra
Introduction
Pyometra, a bacterial infection of the uterus, is diagnosed in up to 23% of intact bitches by 10 years of age.1,2 This hormonally mediated disease occurs in diestrus, during which elevated progesterone concentrations stimulate endometrial hyperplasia and cystic glandular secretion. Reduction in uterine contractility and suppression of local immune defenses increase susceptibility to ascending infection, resulting in the accumulation of purulent intrauterine fluid.3,4 Gram-negative bacterial isolates predominate, with Escherichia coli identified in approximately 66% to 90% of cases.5–7
Common clinical signs include vomiting, polydipsia/polyuria, abdominal pain, and dehydration.8,9 Vaginal discharge may be present, indicating cervical patency.10,11 Anemia, thrombocytopenia, hypoalbuminemia, and azotemia have also been reported.8 Without timely treatment, uterine rupture, septic peritonitis, and death can occur.8,12
While medical management has been described, ovariohysterectomy (OVH), in combination with antimicrobial therapy, remains the treatment of choice for pyometra among both primary and specialty care veterinarians in the US.13,14 In both surgical and medical treatment, broad-spectrum antimicrobials are often prescribed empirically; ampicillin, amoxicillin–clavulanic acid (AMC), enrofloxacin, and a cephalosporin have been recommended.4,15 Antimicrobial selection should be based on expected pathogen susceptibility patterns and subsequently adjusted according to culture and antimicrobial susceptibility testing results.5
Antimicrobial resistance (AMR) in E coli infections has increased in recent years, with multiple veterinary studies reporting resistance to amoxicillin and its derivatives.5,16–18 These findings mirror trends observed in human medicine, in which multidrug resistance (MDR) in E coli has increased significantly over the last decade,19,20 raising concerns regarding ongoing and progressive AMR and reduced antimicrobial efficacy.
Antimicrobial resistance arises from selection pressure associated with antimicrobial use, with overuse and misuse of empiric therapy representing key drivers.19 Geographic variability in AMR has been documented in both human and veterinary medicine, suggesting regional clustering of resistant phenotypes and highlighting the importance of location-specific data to guide empiric therapy and reduce the risk of ineffective treatment.17,21,22 If possible, an empiric antimicrobial agent with at least 80% susceptibility of the anticipated organism should be chosen.23 However, AMR data and disease-specific antibiograms to inform empiric decisions remain limited for many conditions, including canine pyometra.16,17
As veterinary antimicrobial susceptibility testing interpretations published by the Clinical and Laboratory Standards Institute (CLSI) continue to evolve, empiric antimicrobial recommendations for canine pyometra in older veterinary literature may not fully align with current CLSI guidelines. Given the reliance on initial empiric antimicrobial therapy, understanding current antimicrobial susceptibility patterns (ASPs) within the context of evolving CLSI interpretive criteria may help guide more effective treatment.
The primary objective of this study was to describe uterine bacterial culture results and ASPs in dogs with pyometra treated by OVH in the US using a standardized interpretation of raw MIC data. As a secondary objective, we sought to evaluate how evolving CLSI breakpoint criteria influence antimicrobial susceptibility interpretations by comparing the same MIC data for pyometra bacterial isolates using the CLSI VET01S edition in use during the majority of the study period (edition 4)24 and the current edition (edition 7).25 We hypothesized that, when interpreted using CLSI VET01S edition 4, there would be lower susceptibility of E coli to AMC compared to enrofloxacin and that AMC susceptibility would remain below the CLSI-recommended 80% threshold for empiric antimicrobial selection, suggesting it represents a suboptimal empiric antimicrobial choice for canine pyometra. We further hypothesized that reinterpretation of the same MIC data using CLSI VET01S edition 7 would result in lower reported susceptibility to selected antimicrobials, despite no change in the underlying MIC values.
Methods
Case inclusion
Electronic medical record databases of 16 referral hospitals in the US were retrospectively reviewed to identify dogs with pyometra undergoing OVH between January 2020 and December 2024.
Dogs were included if a culture of uterine contents was submitted at the time of OVH for pyometra. If there was no growth on bacterial culture, dogs were included if uterine tissue was submitted for histopathology and changes consistent with pyometra were present. Samples were submitted to reference laboratories according to the individual hospital protocols. The MICs of the antimicrobial susceptibility reports were collected. Dogs with stump pyometra were included if remaining ovarian tissue was present at the time of surgery.
Medical record review
Information obtained from the medical records included breed, age, body weight, clinical signs and duration thereof, length of time since last estrus, and physical examination findings. Specific clinical signs that were tabulated included vomiting, inappetence, lethargy, and polyuria/polydipsia. Physical examination findings recorded included abdominal pain, vaginal discharge, and dehydration level. The results of vaginal cytology, if performed, were noted. Hematological and biochemical abnormalities were recorded. Anemia was defined as Hct or PCV < 35%. Neutrophilia was defined as neutrophil count > 11.64 × 103 cells/µL. Presence of left shift was not assessed owing to incomplete documentation of this parameter in available medical records. Thrombocytopenia was defined as a platelet count < 120 × 103/µL. Azotemia was defined as elevation of serum creatinine (> 1.9 mg/dL) or BUN level (> 27 mg/dL). Hypoalbuminemia was defined as serum albumin < 2.5 g/dL. Using systemic inflammatory response syndrome (SIRS) as a potential marker of sepsis, dogs were scored according to the SIRS criteria described by Spillane et al.26 Dogs meeting ≥ 2 of 4 criteria (temperature, heart rate, respiratory rate, and leukogram abnormalities) were classified as positive for SIRS. Survival to discharge was noted. Length of hospitalization was defined as the number of days from admission to discharge or death for nonsurvivors. If histopathological examination of the uterus was performed, results thereof were included.
Antimicrobial use
Preoperative antimicrobial therapy was defined as antimicrobial administration > 2 hours prior to surgical intervention. Perioperative antimicrobial therapy encompassed the intervening time between 2 hours preoperatively and 24 hours postoperatively. Postoperative antimicrobial therapy was defined as oral antimicrobial therapy administered > 24 hours after surgery and continued after hospital discharge. Type of pre-, peri-, and postoperative antimicrobial treatment and the length of postoperative treatment were recorded.
Antimicrobial susceptibility testing
Bacterial species and raw MIC were extracted from the laboratory antimicrobial susceptibility results. To standardize susceptibility interpretation across laboratories and throughout the study period, all MIC values were retrospectively interpreted using both CLSI VET01S edition 4, which was in use during most of the study period and best reflects the susceptibility information available to clinicians when treatment decisions were made, and the current CLSI VET01S edition 7 breakpoint criteria.24,25 Susceptibility proportions were calculated separately for each edition. Resistance was defined as nonsusceptibility to the antimicrobial. Three laboratories performed antimicrobial susceptibility testing: laboratory A (132 samples; 49 in the Northeast [NE], 50 in the Midwest [MW], and 33 in California [CA]), laboratory B (145 samples; 110 in NE, 28 in MW, and 7 in CA), and laboratory C (1 in MW). Different antimicrobial panels were employed by different laboratories, resulting in varying totals. Median MIC values were calculated for each E coli–antimicrobial combination. Susceptibility percentage was reported as number of susceptible isolates per reported antimicrobial. Regional antibiograms for the most common bacterial isolates were created.
Postoperative antimicrobial prescribing
Postoperative antimicrobial prescriptions were compared with antimicrobial susceptibility results generated with the use of standardized interpretation of MIC values using VET01S edition 4. The same MIC data were subsequently interpreted using VET01S edition 7, evaluating for the influence of evolving breakpoint criteria on isolates nonsusceptible to postoperative antimicrobials.
Statistical analysis
All statistical analyses were performed in Prism, version 10.2.3 (GraphPad Software). Age, clinical signs, number surviving to discharge, number of bacterial isolates, histopathology results, preculture antibiotic use, antimicrobial susceptibility results, postoperative antibiotic prescription, and ASPs by bacteria and region were tabulated. Median MICs were calculated for each organism-antimicrobial combination. Susceptibility proportions were calculated separately using VET01S editions 4 and 7. Age and clinicopathologic findings failed the Shapiro-Wilk test for normality (P < .05) and are reported as medians with ranges. Geographical differences in susceptibility to AMC and enrofloxacin were evaluated using VET01S edition 4. The Fisher exact test was used for regional comparisons. To account for 3 pairwise regional comparisons within each antimicrobial, P values were adjusted with the Holm-Bonferroni method. Adjusted P values are reported.
Results
A total of 666 dogs underwent OVH for suspected pyometra during the study period. Of these, a total of 388 were excluded owing to a lack of bacterial culture (n = 349), combined no growth on bacterial culture and uterine histopathologic changes consistent with pyometra (32), submission of urine culture only (4), culture results not available for review (2), and cancellation of culture due to patient death prior to discharge (1). Two hundred seventy-eight dogs met the inclusion criteria.
Signalment
Sixty-two breeds were represented, of which mixed-breed dogs were most common (n = 76), followed by Pit Bull Terrier (18), Yorkshire Terrier (17), Labrador Retriever (16), and Chihuahua (14). There were 57 additional breeds with 8 or fewer dogs. The median age was 8 years (range, 1 to 16 years), and the median weight was 17.2 kg (range, 1.4 to 69 kg).
Clinical signs
The most common clinical signs, dehydration level, and duration of signs are presented in Table 1. Vaginal discharge consistent with an open cervix was noted in 169 of 278 dogs (60.8%). In-hospital vaginal cytology was performed in 7 dogs, all of which showed neutrophils with intracellular bacteria. Rods were seen in 3 cases, cocci in 1 case, diplococci in 1 case, and both rods and cocci in 1 case. In 1 case, the bacterial morphology was not specified. Systemic inflammatory response syndrome criteria were fulfilled in 253 of 276 dogs (92.0%).
Leia estudo completo:
