Retrospective study on equine uterine fungal isolates and antifungal susceptibility patterns (1999–2011)
Reasons for performing study: Knowledge of commonly encountered fungi infecting the mare's reproductive tract and their respective drug susceptibilities should improve treatment efficacy in mares with fungal endometritis. This is particularly important when practitioners need to start empiric treatment before culture results are complete.
Objective: To report the spectrum of fungal isolates from uterine samples from mares with reproductive problems and their respective antifungal susceptibilities.
Methods: Equine uterine samples submitted to the Cornell University Animal Health Diagnostic Centre for fungal culture between July 1999 and June 2011 were reviewed. Each mare's reproductive history, fungal culture results, antifungal susceptibilities and concurrent aerobic culture results were evaluated. Patterns of antifungal susceptibility and resistance were assessed over time.
Results: One hundred and two fungal isolates were cultured from 92 uterine samples from mares with reproductive problems. Yeast (69%) and mould with septated hyphae (26%) were the most common isolates. Ninety-five to 100% of all fungal isolates were susceptible to the polyenes, while response to the azoles varied with 47–81% of fungal isolates displaying susceptibility. Yeast isolates were 100% susceptible to the polyenes and least susceptible to miconazole (48%) while isolates of mould with septated hyphae were most susceptible to natamycin (100%) and least susceptible to fluconazole (0%). From July 1999 to June 2005 and July 2005 to June 2011, yeast demonstrated increasing resistance to miconazole, while mould with septated hyphae demonstrated increasing resistance to ketoconazole.
Conclusions and clinical relevance: Results from this study suggest that polyenes are effective against uterine fungal isolates in vitro and may be the empiric treatment of choice for fungal endometritis. Isolate resistance to specific azoles increased over time.
The incidence of fungal infection in mares diagnosed with endometritis is significantly lower than the incidence of bacterial infection but fungal endometritis is more challenging to treat and carries a guarded-to-poor prognosis for future fertility [1–12]. Suggested doses and treatment regimens for fungal endometritis are mostly based on case studies or extrapolations from other species (Tables S1–S3). To date, there are no controlled studies on fungal endometritis in mares and pharmacokinetic studies of antifungals rarely include distribution to the reproductive tract.
Antifungal selection to treat fungal endometritis should be based on in vitro susceptibility of specific fungal isolates to available drugs. However, antifungal susceptibility testing is still not offered by many veterinary laboratories and methods are currently being standardised between laboratories by the Clinical and Laboratory Standards Institute. Moreover, in vitro fungal culture requires special media (Sabaroud's dextrose agar or inhibitory mould agar) and prolonged incubation (up to 4 weeks) .
Given the diversity of fungal organisms and paucity of laboratories performing antifungal susceptibility testing, knowledge of the most commonly encountered fungi infecting the mare reproductive tract and their respective drug susceptibilities should help improve treatment efficacy in mares with fungal endometritis. This is particularly important when practitioners need to start empiric treatment before culture results are complete (e.g. when a mare is still in oestrus). To the authors' knowledge, there have been no published studies on the antifungal susceptibilities of uterine-derived fungi from mares. Therefore, the purpose of this study was to report the spectrum of uterine fungal isolates and their respective antifungal susceptibilities, in samples obtained from mares with reproductive problems submitted for fungal culture to the Cornell University Animal Health Diagnostic Centre between July 1999 and June 2011.
Materials and methods
All uterine samples (swabs, biopsies and fluid) submitted to the Cornell University Animal Health Diagnostic Center for fungal culture between July 1999 and June 2011 were reviewed. In this study, inclusion criteria were: 1) samples yielding a positive fungal culture and 2) samples originating from a mare with at least one reported reproductive problem. Reported reproductive problems included vulvar discharge, inflammatory cells detected on endometrial cytology, fluid detected in the uterine lumen on transrectal ultrasound, infertility, endometritis, placentitis or abortion. Antifungal susceptibility profiles and aerobic culture results for submitted samples were also reviewed.
Supplies were purchaseda unless otherwise stated. For fungal culture, samples were inoculated onto both Sabouraud's dextrose agar and inhibitory mould agar. Plates and slants were incubated at 30°C and examined for evidence of growth at least weekly for 4 weeks. Mould identification was based on assessment of macroscopic and microscopic features and evaluation of lactophenol cotton blue mounts [1,2,13,14]. Yeast identification was based on assessment of macroscopic and microscopic features and nutrient assimilation using API 20 C AUX test stripsb or the Uni-Yeast-Tek systemc. Yeast differentiation methods were expanded to include fatty acid profile analysis in 2007 and DNA sequencing in 2009 . The DNA sequencing revealed that many of the genus and species designations assigned to yeast isolates before 2009 were imprecise (Dr C. Altier, personal communication). Therefore, genus designations are only reported for moulds in this study.
Susceptibility testing of the fungal isolates was performed using the Kirby-Bauer disk diffusion method. A standardised inoculum of individual isolates (1:5 dilution of 0.5 McFarland standard suspension) was prepared and plated on DIFCO yeast morphology agar. Plates were dried at 35°C for 10 min and antifungal discsd applied to the agar surface. Antifungal discs included amphotericin B, clotrimazole, fluconazole, itraconazole, ketoconazole, miconazole, natamycin and nystatin. Plates with discs were incubated at 30°C and inspected for fungal growth and zones of inhibition daily for 2 days. Zones of inhibition were measured to the nearest millimetre and used to generate a qualitative report of susceptible, intermediate or resistant (Table S4). For quality control purposes, a known isolate of Candida albicans (ATCC 14053) was tested with every new batch of media.
For aerobic culture, samples were inoculated onto trypticase soy agar supplemented with 5% sheep blood, chocolate agar, Levine EMB agar and Columbia CNA agar supplemented with 5% sheep blood. Plates were incubated at 37°C and examined for growth daily for 2 days. Identification of bacteria was determined by morphological assessment, gram staining, biochemical testing and/or DNA sequencing [1,2,13,14].
Data analysis– Multiple logistic regression models were used to determine if the percentages of isolates susceptible and resistant to individual antifungal drugs and to antifungal classes differed between fungal forms (yeast and mould with septated hyphae, MSH). Antifungal susceptibility data was available for only 1 isolate of mould with nonseptated hyphae (MNH), so this isolate was excluded from analyses. Fisher's exact test was used to evaluate changes in resistance of fungal isolates (grouped by fungal form) to antifungal drugs over time (from July 1999 to June 2005 and July 2005 to June 2011). For evaluation over time, fungal isolates classified as intermediate were grouped with the susceptible isolates. For all analyses, P values <0.05 were considered significant. Data were analysing using JMP version 8 and SAS version 9.2e.
A total of 262 uterine samples were submitted for fungal culture. One-hundred and forty-four (55%) of the samples did not yield fungal growth and 26 (10%) of the samples did not originate from mares with reported reproductive problems. The remaining 92 (35%) uterine samples (83 swabs, 8 fluid samples and 1 biopsy) satisfied the inclusion criteria. Samples originated from mares aged from 5 to 22 years (mean 13.7 years). Breeds were reported for 93% of samples and were mostly Thoroughbred (18), Quarter Horse (17), Warmblood (11), Standardbred (10), American Saddlebred (7) and Arabian (6).
One-hundred and two fungal isolates were cultured from the 92 uterine samples (Table S5). Yeast was the most frequently cultured fungi (70/102, 69%), followed by mould with septated hyphae (primarily Aspergillus spp.; 27/102, 26%). Mould with nonseptated hyphae accounted for the remaining 5% of fungal isolates. The majority (83/92; 90%) of uterine samples yielded one fungal species. Eight samples yielded 2 fungal species, and one sample yielded 3 fungal species.
Fifty-five of the 92 (60%) uterine samples were submitted for aerobic culture in addition to fungal culture (Table S6). Twenty-eight (28/55; 51%) of the samples yielded bacterial growth. The most commonly cultured bacteria were Streptococcus zooepidemicus (9/28; 32%) and E. coli (6/28; 21%). Although all of the samples yielded fungal growth on fungal culture media, only 15 (15/55; 27%) samples yielded growth of fungi on aerobic culture media.
In vitro susceptibility to commonly used antifungals for horses was determined for 74 (74/102; 73%) of the fungal isolates (Table S7). Ninety-five to 100% of fungal isolates were susceptible to the polyenes, while the response to the azoles varied from 47 to 81% susceptibility.
Yeast isolates were 100% susceptible to the polyenes (Table 1). Susceptibilities of yeast isolates to the imidazoles, ketoconazole and clotrimazole were 90 and 84%, respectively. Yeast isolates were significantly more susceptible to clotrimazole and ketoconazole than miconazole (P<0.05), which had the lowest susceptibility (48%). Susceptibility of yeast isolates to fluconazole was also low at 61%.
Table 1. Antifungal susceptibility for yeast isolates from mare uterine samples
|Polyenes|| || || || || || || || || || |
| Amphotericin B||60||60(100)||0||0||24||24(100)||0||36||36(100)||0|
| Natamycin||25||25(100)||0||0||24||24(100)||0||0|| || |
|Imidazoles|| || || || || || || || || || |
|Triazoles|| || || || || || || || || || |
For mould with septated hyphae, 69–100% of isolates were susceptible to the polyenes (Table 2). A significantly lower susceptibility rate to the imidazoles was detected for the mould with septated hyphae isolates compared with the yeast isolates (38–46% vs. 48–90%; P<0.05). Susceptibility of mould with septated hyphae isolates to the triazoles was poor, with 100% of isolates resistant to fluconazole and 54% of isolates displaying resistance or intermediate susceptibility to itraconazole. Antifungal susceptibility was only available for 1 MNH isolate. This Mucor spp. was susceptible to all the antifungals tested (susceptibility to miconazole was not tested).
Table 2. Antifungal susceptibility for mould with septated hyphae isolated from mare uterine samples
|Polyenes|| || || || || || || || || || |
| Amphotericin B||13||9(69)||0||4(31)||5||3(60)||2(40)||8||6(75)||2(25)|
| Natamycin||5||5(100)||0||0||5||5(100)||0||0|| || |
|Imidazoles|| || || || || || || || || || |
|Triazoles|| || || || || || || || || || |
Resistance of yeast isolates to miconoazole increased significantly over time from 0 to 26% (P<0.05, Table 1) whereas resistance to other antifungals did not change significantly. Resistance of isolates from septated moulds increased significantly to ketoconazole from 0 to 63% (P<0.05, Table 2) and there was a trend towards an increase in resistance to itraconazole 0–50% (P = 0.12).
Selection and use of an antifungal based on the in vitro susceptibility of specific fungal isolates to available antifungals is complicated because fungi require special culture media and a prolonged incubation period. This is highlighted in our study by the fact that approximately 73% of positive samples (i.e. fungal growth on Sabouraud's dextrose agar) yielded no growth of fungi on aerobic culture media.
Based on our results, polyenes are a reasonable first choice for treatment of fungal endometritis caused by yeast (100% in vitro susceptibility). Polyenes (amphotericin B, natamycin and nystatin) are organic, broad-spectrum antifungals that bind to ergosterol in the fungal cell membrane, increasing membrane permeability and causing fungal death. The polyenes are water-insoluble and poorly absorbed after oral and local administration, but they have long half-lives after parenteral administration [4,5]. However, polyenes are cytotoxic to mammalian cell membranes, particularly renal tubular cells and vascular smooth muscle cells . Because nephrotoxicity is greatest for natamycin, it is administered only by intra-uterine infusion to avoid systemic absorption . Both nystatin and amphotericin B may be used for intra-uterine infusion or systemic treatment; nystatin may be administered orally and amphotericin B may be administered i.v. [4,5]. To avoid precipitation, amphotericin B must be diluted in sterile water for intra-uterine infusion. For i.v. administration, amphotericin B must be diluted in dextrose and administered slowly to allow monitoring for infusion reactions [4,5]. Lipid formulations of amphotericin B with less potential for adverse reactions have been developed but they are costly and have not been tested extensively in veterinary medicine.
Due to the high susceptibility of yeast isolates to ketoconazole (90%) and clotrimazole (84%), these imidazoles appear to be satisfactory choices for empiric treatment of yeast uterine infections. Azoles are synthetic antifungals that inhibit production of ergosterol, preventing fungal cell membrane synthesis and growth, and are classified as imidazoles (clotrimazole, ketoconazole and miconazole) or triazoles (fluconazole and itraconazole) . Ketoconazole requires dissolution in acidic solution for gastric absorption [4,5]. This makes systemic treatment with ketoconazole impractical because it must be administered via nasogastric tube twice a day to avoid damaging oral and oesophageal mucosa  and concurrent treatment with H2 blockers or protein-pump inhibitors is not recommended . Clotrimazole may be administered by intra-uterine infusion. Adverse effects of the azoles are attributed to impairment of metabolism of endogenous steroids (such as testosterone and cortisol) as well as exogenous drugs [4,5]. Imidazoles have a greater affinity for mammalian cytochrome P450 than the triazoles and are therefore reserved for local treatment (i.e. intra-uterine clotrimazole), while the triazoles are recommended for systemic treatment [4,5]. Fluconazole is the only azole that may be administered intra-uterine, orally or i.v. and it is the preferred triazole because it demonstrates water solubility, minimal protein binding (<10% in man), excellent tissue distribution, has a long half-life (40 h in horses) and few adverse effects [4,5]. In addition, a recent study demonstrated effective endometrial concentrations of fluconazole after daily administration (5 mg/kg bwt per os) . However, due to its widespread use, resistance to fluconazole has been detected in fungal isolates from both veterinary and human patients [7–10]. Similar findings were noted in the current study, as only 60% of yeast isolates were susceptible to fluconazole and yeast isolates were only moderately (76%) susceptible to the other triazole, itraconazole. Susceptibility of yeast isolates was the lowest to miconazole (48%), and yeast resistance to miconazole increased significantly when values obtained from July 1999 to June 2005 were compared with July 2005 to June 2011 (0–26%, P<0.05). Comparatively, these resistance findings are interesting in light of human studies suggesting that the widespread use of azole-based, over-the-counter antifungals to treat vaginitis in women may contribute to selection of resistant strains of yeast .
Selection of an antifungal agent for empiric treatment of uterine infections caused by mould is more challenging than for yeast. Isolates of mould of septated hyphae were the most susceptibile to polyenes (69–100%), but considerably less susceptibile to the imidazoles (38–46%) and the triazoles (0–45%). During the entire study period, mould of septated hyphae isolates were uniformly resistant to fluconazole. When compared with values obtained for isolates from July 1999 to June 2005, isolates obtained from July 2005 to June 2011 were significantly more resistant to ketoconozale (0–63%; P<0.05) and there was a trend towards a similar level of resistance to itraconazole (0–50%; P<0.05 = 0.12). Regardless of sample size, these findings highlight the importance of choosing an antifungal agent for mould infections based on in vitro susceptibility testing.
Polymicrobial infections were present in 51% of samples submitted for both aerobic and fungal culture. The most common uterine bacterial isolates were S. zooepidemicus followed by E. coli, which is in agreement with prior studies . Concurrent fungal and bacterial uterine infections are extremely difficult to treat and require combined systemic antifungal and antibacterial therapy. In addition, intrauterine use of mucolytics (such as acetylcysteine, DMSO and Tris-EDTA; Tables S1–S2) may improve the efficacy of treatment by eliminating biofilm and exposing yeast and bacteria to antimicrobials [15–17].
Future surveillance studies are necessary to determine whether the trend in increasing resistance to antifungal agents continues among fungal isolates. Development of accurate methods for rapid identification of pathogens is critical to avoid reliance on empiric therapy. Although molecular techniques, such as PCR and fluorescence in situ hybridisation are being used more commonly for pathogen identification, they provide no information regarding antimicrobial susceptibility [3,18]. Refinement of these techniques and the addition of susceptibility testing may hold promise for replacement of traditional culture techniques. In the meantime, combination therapy, use of mucolytics and development of antifungals that target novel sites on fungi should be pursued.
In summary, data in this study indicate that the polyenes commonly used to treat equine fungal infections are effective in vitro and may be the empiric treatment of choice for fungal endometritis. Efficacy of fluconazole and miconazole in treating fungal endometritis caused by yeast may be decreased due to an increase in resistance observed in the past 5 years. However, because of its pharmacokinetic properties and safety, fluconazole may be an excellent choice to treat susceptible yeast infections. In our study, susceptibility patterns for isolates of mould with septated hyphae were quite variable and therefore, choice of an antifungal agent for endometritis caused by mould should be based on in vitro susceptibility testing whenever possible. Prospective studies that correlate fungal culture, antifungal susceptibility, treatment and clinical outcome are necessary to verify these results in vivo.
Authors' declaration of interests
No competing interests have been declared.
The authors express profound thanks to Dr. Craig Altier and Ms. Rebecca Franklin for reviewing the Materials and Methods Section and to Dr. Patrick L. McDonough for helpful discussions during study preparation. Authors are also grateful to Dr. Patrick M. McCue for scientific input.
Dr Beltaire assisted with study design and performed most of the data collection, interpretation and preparation of the manuscript. Dr Cheong assisted with study design, data collection and data interpretation; performed data analysis. Dr Coutinho da Silva was responsible for the study design, assisted with data collection, analysis and interpretation and preparation of the manuscript.
a Becton, Dickinson & Co., Sparks, Maryland, USA.
b bioMérieux Inc., Durham, North Carolina, USA.
c Thermo Fisher Scientific Remel Products, Lenexa, Kansas, USA.
d Neo-Sensitabs, Rosco Diagnostics, Taastrup, Denmark.
e SAS Institute, Cary, North Carolina, USA.