The functional competence of uterine-derived polymorphonuclear neutrophils (PMN) from mares resistant and susceptible to chronic uterine infection: a sequential migration analysis

The functional competence of uterine-derived polymorphonuclear neutrophils (PMNs) from 28 mares was measured for migration responsiveness by use of a chamber (filter) assay. Uterine infection was induced with Streptococcus zooepidemicus in mares considered resistant to chronic uterine infection (Grade I). In sequential analysis of uterine flushings obtained from these mares 5, 12, 15, 20, and 25 h after infection was induced, PMNs showed an initial rise at 12 h (from 5), then a general decline in migration response and in concentration of cells per ml from 12 through 25 h post-inoculation. In contrast, PMNs obtained from the uterine flushings from mares considered susceptible to chronic uterine infection (Grade III) demonstrated premature migration dysfunction 12 h after infection. Subsequent increases in functional competence of the PMNs were demonstrated at 15 and again at 25 h after induced infection. The concentration of uterine PMNs per ml from mares considered susceptible to chronic endometritis remained elevated from 12 through 25 h after inoculation, which suggests a possible continued recruitment of new PMNs from the peripheral circulation. The results of this study suggest that uterine-derived PMNs obtained from mares susceptible to chronic uterine infection have a compromised ability to migrate. This dysfunction may play an important role in rendering the endometrium (uterus) susceptible to chronic endometritis.     

Susceptibility to persistent breeding-induced endometritis in the mare: relationship to endometrial biopsy score and age, and variations between seasons

The objectives were to: (1) investigate the associations of age and endometrial biopsy score with uterine fluid retention after insemination; and (2) determine if a strict classification of susceptibility to persistent breeding-induced endometritis (PBIE) based on biopsy score, endometrial cytology, and fluid retention after inseminations, is consistent over subsequent breeding seasons. In Experiment 1, 57 mares were inseminated with 10⁹ freeze-killed sperm during estrus and evaluated for uterine fluid retention 48 h and 96 h after insemination. Comparisons were made between fluid retention and biopsy score or age. In Experiment 2, a subset of 14 mares was classified for susceptibility to persistent breeding-induced endometritis in two subsequent breeding seasons. Biopsy score and age were associated with fluid retention (P < 0.001). In addition, age was related to biopsy score (P < 0.001). Of the mares examined for susceptibility, 36% (5 of 14) changed status during subsequent seasons. Three mares changed to a more severe classification (intermediate to susceptible, or resistant to intermediate), whereas two mares changed to a less severe classification (susceptible to intermediate).     

Effect of susceptibility to endometritis on specific antibody in the endometria of mares

Endometrial biopsies were collected on two occasions from mares resistant to (n = 3) and once from mares susceptible to persistent endometritis (n = 6). The endometrial tissue was minced and cultured in vitro for 24 h. No hemolytic complement activity was detected in the endometrial culture supernatant. Endometrial culture supernatant from mares with persistent endometritis contained titers of antibodies to Streptococcus zooepidemicus similar to those from resistant mares. However, the culture supernatant of biopsies from mares with endometritis was less effective (P < 0.05) at opsonizing S. zooepidemicus in vitro.     

Biodegradable estradiol microspheres do not affect uterine involution or characteristics of postpartum estrus in mares

Quarterhorse mares were used to investigate effects of estradiol-17beta on uterine involution, duration of estrus, interval to ovulation, and fertility achieved by breeding on the first postpartum estrus. On the day of foaling, mares were injected with biodegradable poly (DL-lactide) microspheres containing either 100 mg estradiol-17beta (25 mares) or no drug (27 mares). The treatment period was considered to last for 12 to 15 d. Estrus was determined by teasing mares (n=16) with a stallion. Ovulation was detected by transrectal ultrasonographic examination of ovaries (n=48). On Days 6, 11 and 16 post partum, transrectal ultrasonography was used to measure cross-sectional diameters of the uterine body, uterine horns, and fluid within the uterine lumen (n=28). Uteri were swabbed for bacteriologic culture, and uterine biopsies were obtained from the previously gravid uterine horn on Days 11 and 16 post partum, for assessment of endometritis and morphometric analysis of endometrial histioarchitecture (n=19). Twenty-two mares were bred on foal-heat, and pregnancy was determined by transrectal ultrasonography on 14 to 16 and 30 to 35 d after breeding. With only one exception (diameter of previously gravid uterine horn on Day 11), mean values for all measures of uterine involution did not differ between treatment groups (P > 0.05). No differences were detected between treatment group means for length of estrus or interval to ovulation (P > 0.05). No differences were detected between treatment group liklihoods for recovery of potential bacterial pathogens, presence of endometritis, or presence of intrauterine fluid at 11 or 16 d post partum (P > 0.05). Pregnancy rate of mares treated with estradiol (5 11 ; 45%) was not different from that of control mares (9 11 ; 82%; P > 0.05). Estradiol treatment did not hasten uterine involution, increase duration of estrus, delay ovulation, or increase fertility in these postpartum mares.     

The contribution of complement to opsonic activity in the uterine secretions of mares free of endometritis

The aim of the present study was to investigate if complement contributes to opsonic activity in the uterine secretions of mares with normal reproductive functions. Five mares with a mean age of 9 years were used in the study. The mares were considered to be free of endometritis based upon clinical history, palpation per rectum and ultrasonogaraphy of the genital tract, videoendoscopic inspection of the uterus, electronmicroscopy of endometrial biopsies, and bacteriological and cytological examination of swabs from the endometrium. The hormonal status of the mares was also determined. Uterine secretions were collected by a tampon inserted into the uterus during estrus. Secretion and serum samples were divided in half, one half of which was subjected to heat treatment (56 degrees C in 30 minutes) in order to inactivate the complement. Opsonic activity in the samples was then determined in a chemiluminescence assay using zymosan as the target particle. Opsonic activity was expressed as peak chemiluminescence and the time to peak chemiluminescence. Heat treatment of the secretions reduced peak luminescence significantly (P<0.01), and it prolonged the time to peak luminescence significantly (P<0.05). These data show that complement contributes a considerably to opsonic activity in uterine secretions prior to an inflammatory response in the equine uterus.     

Effects of ecbolic agents on measurements of uterine involution in the mare

Thirteen postparturient mares were used to investigate the effects of ecbolic agents on the rate of uterine involution. Mares were randomly assigned to one of three treatment groups: Group S = intravenous injection of 2 ml saline twice daily for 10 days post partum (n=4); Group O = intravenous injection of 20 units oxytocin twice daily for 10 days post partum (n=4); and Group P = intramuscular injection of 500 mcg fluprostenol twice daily for 10 days post partum (n=5). Ovulation was determined by daily transrectal ultrasonographic examination of the ovaries. On Days 6, 11 and 16 post partum, transrectal ultrasonography was used to measure cross-sectional diameters of the uterine body, uterine horns and fluid within the uterine lumen. Uteri were swabbed for aerobic bacteriologic culture on Days 11 and 16 post partum. Uterine biopsies were obtained from the base of the previously gravid uterine horn on Days 11 and 16 post partum for subjective assessment of endometritis and morphometric analysis of endometrial histoarchitecture. Mean values for all measurements of uterine involution did not differ among groups (P>0.05). For all mares, the diameter of luminal fluid was not correlated to diameter of the uterine body or uterine horns, or to morphometric measurements of endometrial histoarchitecture of the previously gravid uterine horn (P>0.05). Likewise, accumulation of fluid within the uterine lumen was not associated with endometritis or recovery of potential bacterial pathogens (P>0.05). Mean diameter of the previously gravid uterine horn was negatively correlated with morphometric measures of endometrial histoarchitecture of the previously gravid uterine horn (P<0.01).     

Morphologic Evaluation of Acute Endometritis in Mares with Differing Resistance to Uterine Infections

The study was designed to determine differences between normal mares and mares with endometrial pathology in the inflammatory response after bacterial challenge. Six normal mares (biopsy category I) and 4 mares with pathological endometrial changes (biopsy category II) were given an intrauterine infusion of β-hemolytic streptococci on the second day of estrus. All mares had a similar kind of inflammatory response after the bacterial inoculation as assessed by rectal and vaginal examinations. There were no significant differences in the amount of discharge, uterine tone, uterine size and cervical relaxation between the groups. Leukocytic response, as determined by endometrial smears and biopsies, was of the same magnitude in both groups. Two mares from the pathological group were not able to eliminate the infection, but had vaginal discharge and bacteriologically positive uterine swabs until the end of the experiment. It is concluded that the inability of some mares to clear uterine infections cannot be explained by a deficient inflammatory response.     

Use of leukocytes as treatment for endometritis in mares experimentally infected with Streptococcus equi subsp. zooepidemicus

This study compared four treatments for bacterial endometritis in mares experimentally infected with Streptococcus zooepidemicus. Twenty-five mares were used, 20 resistant and five susceptible to endometritis. Mares would be in estrus when infected. Twenty-four hours after inoculation, clinical, bacteriological and cytological examinations were performed and repeated until the first occurrence: negative cytology and no Streptococcus growth or the seventh day post-infection. All mares showed clinical signs of endometritis and were assigned to one of the following treatments: (1) intrauterine infusion of fresh leukocytes; (2) intrauterine infusion of frozen-thawed leukocytes; (3) intrauterine infusion of lysed leukocytes; (4) intrauterine infusion of recombinant human interleukin-8 (rhIL-8); (5) control. Mares were submitted to all treatments, with at least a 14-day interval between treatments in a Latin square design. Treatment did not affect (P=0.121) time needed for resistant mares to eliminate bacteria. Time needed for elimination of bacteria was similar in susceptible mares treated with fresh and frozen leukocytes (P=0.333). Susceptible mares treated with frozen leukocytes also did not differ from those treated with lysed leukocytes (P=0.227) for time to eliminate bacteria, but were significantly different (P>0.02) from those treated with rhIL-8 and control. In resistant mares, physical clearance ability was probably the responsible for bacterial elimination. Intrauterine infusions in susceptible mares with viable or lysed leukocytes associated or not to opsonizing factors, reduced the time to elimination of bacteria. Infusions with bactericidal effect (functional neutrophils and granules) was likely effective and responsible for the more rapid elimination of bacteria in susceptible mares.     

Embryonic loss in pony mares induced by intrauterine infusion of Candida parapsilosis

Pony mares which were detected pregnant by transrectal ultrasonography received a single intrauterine infusion of either sterile saline (control, n = 12 mares) or 10(6)Candida parapsilosis (treated, n = 12 mares) between Days 11 to 14 postovulation. Subsequent embryonic loss was studied by daily ultrasonography of the mare's uterus, by serum progesterone levels, by endometrial swabs for cytologic and microbiologic examination and by endometrial biopsies that were taken after embryonic loss was detected. Significantly fewer (P<0.01) embryonic losses occurred in control than in treated mares (4 12 vs 12 12 ). The mean interval from intrauterine infusion until embryonic loss was 5.8 +/- 2.8 d for control mares (n = 4) and 2.1 +/- 0.2 d for treated mares (n = 12). Prior to embryonic loss, moderate to marked edema of the endometrial folds in 12 of 12 treated mares and free fluid in the uterine lumen of 5 of 12 treated mares were detected by ultrasonography. After embryonic loss, Candida parapsilosis was cultured from the uteri of 8 of 12 treated mares, and E . coli was cultured from the uteri of 2 of 4 control mares. Postloss endometrial smears had cytologic evidence of inflammation in 10 of 12 treated mares and 3 of 4 control mares. Intrauterine inoculation of C. parapsilosis consistently induced embryonic loss and may provide a basis to further study the relationship between endometritis and embryonic loss in mares.     

Lysozyme, alkaline phosphatase and neutrophils in uterine secretions of mares with differing resistance to endometritis

A study was conducted to 1) determine differences in the inflammatory response following bacterial challenge between normal mares and mares with chronic endometritis and 2) to determine if enzyme activity in uterine fluid can be used to evaluate degree of inflammation in the equine uterus. Six normal mares (Group 1) and four mares with chronic endometritis (Group 2) received an intrauterine infusion of beta-hemolytic streptococci on the second day of estrus. Neutrophil concentration as well as lysozyme and alkaline phosphatase activity were determined in uterine secretions obtained by placing tampons in the uterus of mares. All mares had a similar inflammatory response following bacterial challenge of the uterus, as indicated by a neutrophil response of the same magnitude. Neutrophil numbers, lysozyme and alkaline phosphatase concentrations were all increased 12 h postinoculation and declined rapidly to normal preinoculation values by 48 h after inoculation. In spite of the similarity of the clinical signs, neutrophil concentrations and enzyme activity, mares in group 1 demonstrated a markedly higher ability to eliminate the infection than mares in group 2. It is concluded that factors other than neutrophil numbers, lysozyme and alkaline phosphatase activity account for the inability of the mare to eliminate uterine infections.     

The effect of transcervical uterine manipulations on establishment of uterine infection in mares under the influence of progesterone

Four pony mares were used in a cross-over study to investigate the effect of different treatments on experimentally-induced endometritis. The mares were treated with progesterone to facilitate establishment of uterine infections. They received an intrauterine infusion of Streptococcus zooepidemicus 5 days after the start of progesterone therapy. Five days later, they were treated by intrauterine infusions of 2 g ampicillin in 50 ml sterile water or by sterile water without antibiotic for 3 consecutive days. Prior to infusion of Strep. zooepidemicus , no bacteria were cultured from the uteri of the mares. However, 5 days after infusion of Strep. zooepidemicus and prior to antibiotic therapy, mixed bacterial growths were cultured from endometrial swabbings. After antibiotic therapy, ampicillin-resistant organisms were cultured from endometrial swabbings. Two other progesterone-treated mares received an intrauterine infusion of sterile phosphate buffered saline instead of bacteria. Mixed bacterial cultures were recovered 5 days later from the endometrial swabbings of these mares. It was concluded that the high circulating concentrations of progesterone were probably responsible for the treatment failure and that in clinical situations, therapy involving transcervical manipulations should not be administered when mares are in diestrus.     

Does artificial insemination with chilled, extended semen reduce the antigenic challenge to the mare's uterus compared with natural service?

Uterine response in infection-resistant mares (n = 5) at 48 h after AI was compared with that following natural service in these same mares, and after AI in infection-susceptible mares (n = 6). In the resistant mares, small amounts of uterine fluid were detected and bacteria were isolated infrequently at 48 h after breeding, but cytological examination of uterine flushes revealed that a significant degree of endometritis was present. There was no difference in the degree of inflammatory response by 48 h after AI or natural mating. In 4 of the 6 susceptible mares moderate to large accumulations of intrauterine fluid were detected at 48 h after AI, and massive uterine neutrophilia was present in all 6 mares. It was concluded that there was no evidence that using conventional AI techniques limited the inflammatory response of the mare's uterus.     

Effect of oxytocin and flunixin meglumine on uterine response to insemination in mares

The most probable reason for persistent postbreeding endometritis in mares is weak myometrial contractility. The influence of oxytocin (OT; an ecbolic agent) and flunixin meglumine (FLU; a prostaglandin inhibitor serving as a model for mares with decreased uterine contractility) on uterine response to artificial insemination (AI) was studied in mares with no history of reproductive failure. The mares were treated intravenously with 10 mL saline (Group C, n = 10) or 0.01 IU/kg OT (Group OT, n = 10) 2, 4, 8, and 25 h after AI. Group FLU (n = 11) was treated with 1.1 mg/kg FLU 2 h after AI and with saline thereafter. The mares received the same treatments in the first and third cycles but were sampled either at 8 or 25 h. The amount of intrauterine fluid (IUF) and edema and the number of uterine contractions were recorded before AI and 10 min after the treatments using transrectal ultrasonography. At 8 h after AI, the mares were treated with human chorionic gonadotropin, and, after 8-h or 25-h scans, a 500-mL uterine lavage and a biopsy were performed. Ovulation was confirmed at 48 h and pregnancy 14 to 17 d after AI. No manipulations were done during the second estrus. At 8 h after AI, Group FLU had more polymorphonuclear leukocytes (PMNs) in the uterine lavage fluid than did Group OT (P < 0.05), but uterine contractions did not differ significantly. At 25 h, the PMN concentrations were low in all groups. Group OT rarely showed IUF. The uterine biopsy specimens of Group FLU showed less inflammation of the stroma but more PMNs in the uterine lumen 8 h after AI than that of the control group (P < 0.05). The pregnancy rates did not differ between the groups (63% C, 53% OT, and 50% FLU). Oxytocin rapidly and effectively removed IUF and PMNs after AI and thereby shortened the duration of postbreeding inflammation.     

Differences in uterine position of reproductively normal mares and those with delayed uterine clearance detected by scintigraphy

The position of the uterus within the abdomen may affect a mare's ability to rapidly clear the uterine lumen of contamination. In this study, the position of the uterus was determined from left and right lateral flank scintigrams taken 1 and 2 h after intrauterine infusion of radiocolloid. Scintigraphy was performed during estrus in 44 mares, 24 were reproductively normal and 20 exhibited a delay in uterine clearance. Reproductively normal mares were nulliparous (n = 14) or pluriparous (n = 10), 3 to 21 yr of age, had no history of persistent uterine infections and cleared > 50% of a radiocolloid within 2 h of infusion into the uterus. Mares that exhibited a delay in uterine clearance were pluriparous (n = 18) or nulliparous (n = 2), 12 to 24 yr of age, had a history of endometritis and cleared < 30% of a radiocolloid within 2 h. The angle between the caudal-ventral aspect of the uterine image and cervix relative to horizontal as visualized on the scintigram was measured with a protractor. Results were analyzed by the General Linear Model System. The uterine-cervical angle relative to horizontal was more ventral in mares with delay in uterine clearance and was more horizontal in reproductively normal mares (mean +/- SEM-111.6 +/- 3.6 for delay in uterine clearance mares; 147.6 +/- 3.9 for reproductively normal mares; P < 0.0001). The mean angle for reproductively normal, pluriparous mares was steeper than that for nulliparous mares (141.1 +/- 2.9, 152.3 +/- 2.44 respectively; P = 0.004). There were no differences in angles between left and right lateral views within individuals. We conclude that a uterus that tilts ventrally in relation to the pelvic brim may contribute to the inability of delay in uterine clearance mares to rapidly clear their uterine lumen of contamination. Parity may contribute to the more ventral orientation of the uterus.     

Comparison of three diagnostic methods to identify subclinical endometritis in mares

The objective of this study was to compare the accuracy of a uterine swab (US), a cytological brush (CB) and an endometrial biopsy (EB) to detect subclinical endometritis in mares. Cytological and bacteriological results of all three techniques were related to histological occurrence of polymorphonuclear neutrophils (PMNs) in the stratum compactum, commonly known as ‘best standard'; to diagnose endometritis.Samples were taken from 55 mares of different breeds without clinical signs of endometritis. Samples for US, CB and EB were collected, smeared on a microscopic slide and cultured for bacterial growth. Endometrial biopsy samples were additionally stored in 4% formaldehyde for histological analysis. Bacteriological cultures and cytological samples of all techniques were classified as negative (no uterine pathogens in monoculture; < 2% PMNs) or positive (uterine pathogens in > 90% of the grown colonies; > 2% PMNs) for endometritis. Uterine pathogens were diagnosed in 20.0% of the mares. Isolation of pathogens was not associated with positive cytological findings (r = −0.23; P = 0.87). None of the six mares with an Escherichia coli infection (10.9%) showed a positive cytological result. In contrast, two of five mares infected with Streptococcus zooepidemicus had a positive cytological result.Histologically, the presence of PMNs in the stratum compactum was regarded as positive for endometritis when the mare was in diestrus at time of sampling. Compared to the ‘best standard', sensitivity for cytology of CB, US and EB was 0.17, 0.00 and 0.25, respectively. Specificity for cytology of CB, US and EB was 0.83, 0.93 and 0.85, respectively. Sensitivity of uterine culture was 0.25, 0.33 and 0.25 for CB, US and EB, respectively. Specificity for culture of CB, US and EB was 0.80, 0.83 and 0.95, respectively. In conclusion, cytological or bacteriological examinations alone provide a high incidence of false negative results. Sensitivity of cytology combined with bacteriology of CB was 0.42. A combination of a bacteriological and a cytological examination of a CB sample improved the diagnostic performance in subfertile mares. Based on these results, we can recommend the CB to improve the diagnosis of subclinical endometritis in the mare compared to the US alone as currently used routine method.     

Immunoglobulin levels,protein concentrations and alkaline phosphatase activity in uterine flushings from mares with endometritis

Uterine samples and flushings were obtained from 87 mares to compare endometrial bacteriology and biopsy with immunoglobulin and protein concentration and alkaline phosphatase activity in uterine flushings. Mares were designated as infected if both bacteriology and biopsies were positive. The immunoglobulin levels, protein concentration and alkaline phosphatase activity in uterine flushings from infected and non-infected mares were compared. Twenty (23%) of the mares were classified as infected. A significantly higher proportion of infected mares (cf. non-infected) had elevated IgA and protein concentrations. Levels of IgG, IgG_T or alkaline phosphatase were not significantly elevated in infected mares. These results suggest that IgA and protein levels are elevated in the uterus in the presence of active infection.     

Uterine vascular degeneration is present throughout the uterine wall of multiparous mares. Colinearity between elastosis, endometrial grade, age and parity

Vascular degeneration is present in endometrial vessels of multiparous aged mares. The lesions associated with vascular degeneration consist of enlargement, duplication and splitting of the membrana elastica interna and perivascular deposits of elastin. However, there are no similar data available for deep myometrial vessels and the vascular layer. The objectives of the present study were to characterize the status of vasculature in full-thickness uterine necropsy samples and to correlate these findings to endometrial grade, age, and parity. Elastosis was present in myometrial vessels, as well as in large arteries and veins located between the circular and longitudinal myometrial layers. Vascular degeneration was associated with number of foals (P < 0.001) and endometrial grade (P < 0.05), but not with mare age (P > 0.05). Endometrial grade was associated with age (P < 0.001) and vascular grade (P < 0.05), but not with number of foals (P > 0.05). The presence of elastosis in the myometrial vessels was related to problems associated with chronic uterine infection (CUI) and delayed uterine clearance (DUC) of infertile mares. Uterine contractility was impaired in mares affected by CUI and/or DUC and could be related to a lack of myometrial blood flow. Additionally, degeneration of large vessels in the vascular layer may indicate a general compromise in uterine blood flow and fertility. The main conclusions were the presence of vascular elastosis in large deep myometrial vessels as well as in endometrial vessels, and that the factor with the strongest association with vascular degeneration was number of foals (P < 0.001), followed by endometrial grade (P < 0.05), but no association with mare age.     

Mares with delayed uterine clearance have an intrinsic defect in myometrial function

Persistent, postmating endometritis affects approximately 15% of mares and results in reduced fertility and sizable economic losses to the horse-breeding industry. Mares that are susceptible to postmating endometritis have delayed uterine clearance associated with reduced uterine contractility. Unfortunately, the mechanism for reduced uterine contractility remains an enigma. The present study examined the hypothesis that mares with delayed uterine clearance have an intrinsic contractile defect of the myometrium. Myometrial contractility was evaluated in vitro by measuring isometric tension generated by longitudinal and circular uterine muscle strips in response to KCl, oxytocin, and prostaglandin F(2alpha) (PGF(2alpha)) for young nulliparous mares, older reproductively normal mares, and older mares with delayed uterine clearance. In addition, intracellular Ca(2+) regulation was evaluated using laser cytometry to measure oxytocin-stimulated intracellular Ca(2+) transients of myometrial cells loaded with a Ca(2+)-sensitive fluorescent dye, fluo-4. For all contractile agonists, myometrium from mares with delayed uterine clearance failed to generate as much tension as myometrium from older normal mares. Oxytocin-stimulated intracellular Ca(2+) transients were similar for myometrial cells from mares with delayed uterine clearance and from older normal mares, suggesting that the contractile defect did not result from altered regulation of intracellular Ca(2+) concentration. Furthermore, no apparent age-dependent decline was observed in myometrial contractility; KCl-depolarized and oxytocin-stimulated longitudinal myometrium from young normal mares and older normal mares generated similar responses. However, circular myometrium from young normal mares failed to generate as much tension as myometrium from older normal mares when stimulated with oxytocin or PGF(2alpha), suggesting possible age-related alterations in receptor-second messenger signaling mechanisms downstream of intracellular Ca(2+) release. In summary, for mares with delayed uterine clearance, an intrinsic contractile defect of the myometrium may contribute to reduced uterine contractility following breeding.     

Use of a low-volume uterine flush for diagnosing endometritis in chronically infertile mares

Low-volume uterine flush (n=401) was performed in 308 infertile mares to diagnose endometritis. Mares evaluated were either barren after three or more breedings or had two or more unsuccessful embryo recovery attempts during consecutive cycles. Culture results were compared with cytological and histological findings, efflux clarity and pH to substantiate that the micro-organisms recovered were truly pathogens. Cytological specimens were evaluated for presence of epithelial and inflammatory cells, bacteria, yeast and debris. Endometrial biopsies (n=110) were examined for the presence of neutrophils in the stratum compactum. Micro-organisms were recovered in 282/401 (70%) of low-volume flushes; E. coli was most frequently isolated (42.2%), followed by beta hemolytic Streptococcus (37.6%). Efflux clarity of 318 flushes was clear (n=109), cloudy (n=149), or mucoid (n=60). Isolation of micro-organisms was highly associated with cloudy and mucoid effluxes (P<0.001), debris on cytological specimens (P<0.001), increased efflux pH (P<0.003), and neutrophils on endometrial biopsy (P<0.01). E. coli was associated with debris on cytological smear (P<0.002), whereas beta hemolytic Streptococcus was associated with increased efflux pH (P<0.002). Using the presence of neutrophils in a tissue specimen as the "best standard" for diagnosing endometritis, the sensitivity of flush culture was 0.71 and for flush cytology was 0.8, whereas the specificity was 0.86 and 0.67, respectively. Neutrophils in uterine flushes under-reported inflammation; only 86/282 positive cultures were positive on cytology. The clinical estimate of a contaminated (false positive) flush culture was 11%, if a false positive was defined as positive culture with clear efflux and no debris or neutrophils on cytology (26/228). In conclusion, a low-volume uterine flush was a rapid, accurate method for identifying mares with chronic endometritis. When micro-organisms were recovered, endometritis was confirmed by efflux clarity, pH and cytological findings of debris, bacteria, or neutrophils. E. coli was most commonly isolated and it appeared to differ in pathogenicity from beta hemolytic Streptococcus.