Early, accurate diagnosis of ascending placentitis in mares remains a key challenge for successful treatment of the disease. Doppler ultrasonography has shown promise as a tool to diagnose pregnancy abnormalities and is becoming more available to equine clinicians. However, to date, no studies have prospectively compared this technique to standard B-mode measurement of the combined thickness of the uterus and placenta (CTUP).
The objective of the current study was to compare Doppler and B-mode ultrasonography for the detection of experimentally-induced ascending placentitis in mares.
Eleven healthy pony mares in late gestation were used in this study. Placentitis was induced in 6 mares between Days 280 and 295, while 5 mares served as negative controls. All mares were intensively monitored until delivery. Fetal heart rate, CTUP, uterine artery blood flow (resistance index, pulsatility index, arterial diameter and total arterial blood flow) and physical examination findings were recorded at each examination. Mares with an increased CTUP above published values were treated in accordance with published recommendations. Foals and fetal membranes were examined at birth. Ultrasonographic parameters were compared between groups using ANOVA. Foal viability and histological presence of placentitis were compared using a Fisher's exact test.
The CTUP was increased above normal in 5 of 6 inoculated mares within 3 days after inoculation (P = 0.05). The sixth inoculated mare was excluded from subsequent data analysis. Uterine artery blood flow, physical examination findings and fetal heart rate were not different between groups. Gradual increases in CTUP, arterial diameter and total arterial blood flow were detected with increasing gestational age in the control mares (P = 0.02, P = 0.00001 and P = 0.00001, respectively).
The CTUP, but not uterine blood flow, was different between groups (P = 0.00001). Recorded CTUP values for control pony mares were similar to previously published values for light breed horses.
Ascending placentitis is the most important infectious cause of abortion and neonatal loss in horses. It markedly affects the equine breeding industry with an estimated impact of $2.2 billion nationally, according to the 2005 National Economic Impact Study of the American Horse Council Foundation . Giles and colleagues examined aborted fetuses and placental tissues from 3527 animals and determined that placental pathology was responsible for about one-third of the late-term fetal losses and neonatal deaths [2, 3].
Placental infection in mares often results in rapid abortion or premature delivery without premonitory signs [4, 5]. However, if premature labour is delayed, a foal may precociously mature and be born viable [6, 7]. An important limitation to successful treatment outcome is the challenge of timely diagnosis of disease before irreversible placental or fetal damage has occurred. A second limitation to directed treatment of mares with placentitis is the lack of effective tools to monitor the disease after diagnosis. Treatments are generally continued empirically until foaling, abortion or ultrasonographic evidence of fetal death [8, 9]. In other cases, treatments are discontinued after a period of several weeks or given intermittently over the course of gestation. Identification of a diagnostic tool that could be used to monitor treatment effect could help guide clinical decision-making during the course of therapy and result in directed treatment of disease.
Advances in the technology available to equine clinicians worldwide represent an opportunity to improve the accuracy of identifying placentitis and other diseases. Ultrasonographic measurement of the combined thickness of the uterus and placenta (CTUP) is currently the most widely used diagnostic tool for diagnosis of placentitis [9-11]. To date, no study has compared this modality to newer technologies, such as Doppler ultrasonography. Doppler ultrasonography is now widely available in portable ultrasound units and may be a sensitive tool for detecting disease. Specifically, 3 measures of blood flow – the resistance index, pulsatility index and total arterial blood flow have been most commonly applied to quantify blood flow via Doppler ultrasonography [12-16]. The resistance index and pulsatility index are indirect measures of blood flow to an organ or tissue that account for the resistance in flow to that organ, while total arterial blood flow measures blood flow within a single cardiac cycle at a specific location. Doppler ultrasonography is a well-established diagnostic technique in human medicine and obstetrics [17-21]. It is noninvasive, can be rapidly performed and is a valuable predictor of pre-eclampsia and intrauterine growth retardation well before clinical signs of disease are evident. Studies in the horse have established normal values for blood flow in the uterine arteries of the nongravid and gravid horns throughout pregnancy [12, 22-24]. Significant changes in blood flow of the ipsilateral and contralateral uterine arteries have been reported at the time of experimentally induced embryonic death in the horse .
The objective of the current study was to compare the ability of B-mode and Doppler ultrasonography to detect ascending placentitis in mares with experimentally induced disease and to monitor progress of treated mares with a known infection. We hypothesised that estimating uterine artery blood flow via resistance index, pulsatility index and total arterial blood flow utilising Doppler ultrasound technology would result in early detection of placentitis in experimentally infected pony mares. Specifically, we hypothesised that the resistance index and pulsatility index would decrease, while total arterial blood flow would increase after inoculation and prior to changes in the CTUP.
Materials and methods
Eleven pregnant pony mares were enrolled in the study between 250 and 270 days of gestation between January 2011 and June 2011. Mares were between 2 and 12 years of age (mean age 7.2 years) and weighed between 166.5 and 511.5 kg (mean bwt 345.5 kg). All mares were a part of an established research herd. Each mare received a complete breeding soundness examination upon arrival and all mares were bred at the research facility. Mares were housed in grass paddocks at the NCSU Equine Health Center at Southern Pines and received standard preventative health maintenance as dictated by the recommendations of the American Association of Equine Practitioners and University policy and in accordance with institutional IACUC policies.
Induction of ascending placentitis
Between 280 and 295 days of gestation, mares were randomly divided into 2 groups. Six mares were inoculated intra-cervically, while 5 mares, matched by gestational age, served as negative controls. Inoculated mares were infected with 107 colony forming units of Streptococcus equi subsp. zooepidemicus (S. zooepidemicus) obtained from a stock solution prepared from a clinical isolate submitted to the Microbiology Laboratory at the University of Florida, College of Veterinary Medicine in 1999 . A stock solution of the bacterial isolate was stored in cryovials containing Brucella broth with 10% glycerol and porous beads (Cryosaver)1 at -80°C. The bacterial isolate was sensitive to trimethoprim sulfamethoxazole in vitro. Two days prior to inoculation, an aliquot was thawed and replated to confirm a pure culture. On the day of inoculation, a 107 colony forming unit inoculate was made using the McFarland standards for microbiology dilutions (McFarland Standard 0.5)1 and diluted in 1 ml 0.9% saline immediately prior to inoculation. Bacteria were deposited midway through the cervix using digital guidance, in accordance with published methodology [4, 27]. Control mares had no vaginal intervention (i.e. sham inoculation) but were examined on the same schedule as their infected counterparts.
All mares were examined by transrectal and transabdominal ultrasonography twice weekly from the time of enrolment until inoculation. To perform the examinations, mares were restrained in stocks modified to safely accommodate ponies. Following inoculation, all mares were examined ultrasonographically every 12 h for 5 days, daily for 8 days and then every other day until foaling or abortion had occurred. Three measurements of fetal heart rates were obtained via transabdominal ultrasound using a 3–5 MHz curvilinear ultrasound transducer (Sonosite MicroMaxx)2 .
A 5–7 MHz linear transducer (Sonosite MicroMaxx)2 was used for Doppler ultrasonography of the uterine arteries. The transducer was placed against the lateral aspect of the caudal aorta in the mid-abdomen and retracted gradually to the origin of the external iliac artery (Fig 1a). At this point, the external iliac artery was traced until both the uterine artery and deep circumflex artery were identified (Fig 1b). The diameter of the uterine artery was measured at the level that it crossed the deep circumflex artery or approximately 2–5 cm distal from its origin. Colour Doppler ultrasound was used, as previously described, to obtain blood flow measurements of the uterine arteries [13, 16, 25, 29]. Briefly, a longitudinal section of the artery was identified by retroflexing the transducer slightly to face dorsomedially (Fig 1c) and the Doppler gate was placed within the artery for measurement. After completion of measurements on one side, the arm used for examination was switched to obtain measurements on the contralateral side (i.e. the right arm was used for the left artery and the left arm for the right artery). The pulse wave colour Doppler ultrasound had a sample gate setting of 3 mm and angle correction was set between 30 and 60°. Three representative waveforms were obtained and spectral analysis performed using the algorithm package provided with the ultrasound unit to determine resistance index, pulsatility index and the maximum velocity (Vmax) (Fig 2). These known parameters were later used to calculate total arterial blood flow (TABF = [(Vmax + (RI * Vmax - Vmax))/PI] * (D/2)2π), where RI = resistance index, PI = pulsatility index and D = arterial diameter [15, 30]. At the conclusion of the examination, 3 measurements of the CTUP were obtained as described previously [9-11]. Measurements were taken just cranioventral to the cervix at a location where minimal folding of the uterus and placenta was present (Fig 3). Measurements were taken no closer than 1–3 cm from each other. In cases where separation of the uterus and placenta were noted rather than thickening, the caliper was placed across the widest point of separation. For each parameter, 3 measurements were obtained and averaged for statistical analyses.
Treatment was initiated in inoculated mares after CTUP values were detected that exceeded the published normal ranges . Maximum normal measurements of CTUP in light horse breeds were previously described by determining the mean thickness of the uteroplacental unit during 4 time periods (<270 days, 270–300 days, 300–330 days and >330 days of gestation). The mean thickness plus 2 s.ds was classified as the maximum normal thickness for CTUP for gestational age: 7 mm for <270 days, 8 mm for 270–300 days, 10 mm for 300–330 days and 12 mm for >330 days [9, 10]. In the current study, mares were therefore classified as having ultrasonographically detectable disease when the mean CTUP from 3 measurements exceeded the published normal value for gestational age. In order to approximate the therapeutic management of placentitis in a clinical setting, drugs were administered simultaneously at doses consistent with those used clinically. Trimethoprim sulfamethoxazole3 (30 mg/kg bwt per os q.12 h), altrenogest4 (0.088 mg/kg bwt, per os q. 24 h) and pentoxifylline5 (8.5 mg/kg bwt, per os q. 12 h) were administered to mares until abortion, stillbirth or delivery of a live foal.
In addition to the scheduled ultrasound examinations, all mares underwent a daily physical examination during which pulse, respiratory rate and temperature were recorded. Physical changes consistent with foaling or placentitis, including mammary development, lengthening of the vulva and laxity of the tailhead and gluteal region, were noted. Additionally, mares were monitored for foaling using an electronic foaling sensor (Foalert system)6. A transmitter was sutured across the vulva of each mare either the day after inoculation or 2 weeks prior to the expected foaling date (gestational Day 320). At this time, mares were moved to a round pen within the paddock to prevent herd mates from attempting to adopt the newborn. Periparturient mares were monitored by frequent observation in the paddock every 4–6 h. When evidence of parturition was noted visually (increased incidence of recumbency, restlessness, inappetence, straining to urinate or evidence of fluid from the vulva indicating rupture of the chorioallantois) or when the system signalled separation of the vulvar lips, mares were immediately placed under continuous observation. Mares were allowed to foal in the paddock without intervention unless clinically indicated.
At delivery, foaling was classified as abortion (premature delivery of a dead foal prior to 320 days), stillbirth (delivery of a dead full-term foal between 320 and 360 days), delivery of a nonviable foal, or delivery of a viable foal. All live foals immediately underwent a physical examination. Foals that were able to breathe without mechanical assistance, right themselves after birth, respond to nasal or ear stimulation and had good muscle tone were deemed viable. Viable foals were allowed to bond with mares in the immediate post partum period with minimal intervention. Live, nonviable foals were subjected to euthanasia by intravascular injection of pentobarbital (Beuthanasia)4.
Tissue samples of fetal membranes were collected immediately after passage of the membranes. Five centimetre pieces of tissue were taken from the chorioallantois (cervical star region/body, pregnant horn and nonpregnant horn), amnion and mid-umbilicus. Any grossly abnormal area of the fetal membranes was also sampled. All samples were placed in formalin and processed by the clinical histopathology laboratory at North Carolina State University. Stained slides were evaluated by a pathologist who was blinded to the status of the animals for presence of bacteria, inflammation and tissue necrosis. Animals were classified as normal or demonstrating chorioallantoic inflammation, amniotic inflammation and funisitis.
Based on published data of Doppler ultrasonography in normal pregnant mares, a 20% s.d. is observed during late gestation [12, 31]. In order to detect a 35% difference with a type I error of 0.05 and the probability of a type II error of 0.8, sample size was determined to be a minimum of 5 animals per group via a priori analysis using a power analysis programme (G*Power 3.1)7 [32, 33]. Statistical analysis of the data was performed using a commercial software package (Statistix 8.1)8. Foal viability and histological presence of placentitis were compared between groups using a Fisher's exact test. Gestational age at foaling was compared between groups using a 2 sample t test. Mean daily values for CTUP, fetal heart rate, resistance index, pulsatility index, arterial diameter and total arterial blood flow were tabulated for analysis. For resistance index, pulsatility index, arterial diameter and total arterial blood flow, measurements of both arteries were averaged together. Data were analysed for normality using a Shapiro–Wilk test. A log transformation was used for data with a non-normal distribution (fetal heart rate, resistance index, pulsatility index and total arterial blood flow). Measurements for both groups were blocked by gestational age (250–270, 270–300, 300–330, >330) and analysed for differences using ANOVA. Mean daily values for each parameter from 3 preinoculation days and the first 11 days post inoculation were compared between groups. Mean daily values for measures taken 0–1, 2–3, 4–5 and 6–7 days prior to foaling were compared between groups and between those mares delivering viable and nonviable foals, in order to detect differences in the immediate prepartum period. An ANOVA with repeated measures test was used for these analyses. A least significant difference (LSD) all pairwise comparisons test was used to compare means among days. A significance level of P<0.05 was established. A tendency toward significance was reported for P>0.05 and <0.1. Numerical data were reported in the text and tables as mean ± s.d.
Five of 6 inoculated mares were diagnosed with placentitis based on increased measurements of the CTUP within 3 days after inoculation [9, 10]. The sixth mare failed to develop an increased CTUP within 10 days of inoculation; therefore data from that mare were excluded from analysis of ultrasonographic measurements. Ultrasound data from 306 examination days from 5 infected mares and 5 control mares were included in the analysis.
Two of 5 infected mares delivered viable foals, one mare had a stillbirth and 2 foals were born alive but nonviable. All control mares delivered live, viable foals (P = 0.08). Physical examination parameters (temperature, pulse, respiration) were not different between groups at any time during the experiment (P>0.4). Infected mares had a shorter gestation length compared with the control group, (309 ± 11, range; 300–326 and. 332 ± 14, range: 316–351 days, respectively; P = 0.02). Mares delivering nonviable foals delivered significantly earlier than mares delivering viable foals in either group (301 ± 1.5, range: 300–303 and 328 ± 13, range: 315–351 days, respectively; P = 0.009). The inoculated mare that failed to develop an increased CTUP delivered a live foal at 351 days of gestation.
In control mares, CTUP increased in a sequential fashion with gestational age between 250 and 330 days, as described previously for horse mares , but no further increase was detected after 330 days (P = 0.02; Table 1). In infected mares, the gestational changes were obscured by a large increase in thickness after inoculation (Table 1). The CTUP measures were higher in infected mares than control mares beginning 2 days after inoculation (1.30 ± 0.56 and 0.61 ± 0.11 cm, respectively; P = 0.00001) and tended to be higher until foaling (1.31 ± 0.68 and 0.56 ± 0.10 cm, respectively; P = 0.07) (Fig 4). Placentitis was diagnosed in 5 inoculated mares based on CTUP exceeding published normal values by Day 3 after inoculation (Fig 4). The CTUP remained increased in 4 of 5 mares until foaling despite long-term treatment. In one mare, CTUP returned to within normal limits 15 days after treatment onset and remained within normal limits until foaling 16 days later (gestational Day 326). None of the control mares had an CTUP that exceeded published values for light horse breeds  (Fig 4).
Table 1. Findings of B-mode and Doppler ultrasonography during late gestation in pony mares with normal pregnancies and mares with experimentally induced ascending placentitis
Group CONT (n = 19)
Group INOC (n = 12)
Group CONT (n = 72)
Group INOC (n = 81)
Group CONT (n = 74)
Group INOC (n = 26)
Group CONT (n = 19)
Group INOC (n = 0)
Data for CTUP (combined thickness of the uterus and placenta), fetal heart rates (FHRs), mean diameter of the uterine arteries (D), resistance index (RI) pulsatility index (PI) and total arterial blood flow (TABF) are reported as mean ± s.d. for 4 time periods in late gestation and grouped based on inoculation status. Data were analysed for normality with a Shapiro–Wilk test. Data with a non-normal distribution were log-transformed. Differences between groups were detected using ANOVA. Different uppercase superscripts (A,B,C) indicate differences between gestational age blocks in control mares (P<0.05). Different lowercase superscripts (a,b,c) indicate differences between gestational age blocks in infected mares (P<0.05). Different numbers (1,2) indicate differences between groups within a gestational age (P<0.05). INOC = inoculated; CONT = Control.
Fetal heart rate in both groups also followed the previously described pattern for normal gestation [34, 35], with a gradual decline in rates between 250 and parturition (P = 0.00001; Table 1). Fetal heart rate were not different between groups (P = 0.3; Table 1). At the last measurement prior to foaling, fetal heart rates were not different for viable and nonviable foals (86.1 ± 21.7 and 76.2 ± 7.7 beats/min, respectively; P = 0.3).
In both groups, arterial diameter (infected mares, P = 0.007 and control mares, P = 0.00001) and total arterial blood flow (infected mares, P = 0.017 and control mares, P = 0.00001) increased with advancing gestational age (Table 1). The flow indices (resistance and pulsatility) were not different between gestational ages in either group (Table 1). Doppler measures were not different between groups at any time point after inoculation. At the last measurement prior to foaling, Doppler measures did not differ between mares that delivered viable or nonviable fetuses (arterial diameter: 0.83 ± 0.07 and 0.71 ± 0.15 mm, respectively; total arterial blood flow: 334 ± 120 and 256 ± 99 ml/min, respectively; resistance index: 0.51 ± 0.1 and 0.45 ± 0.02, respectively; pulsatility index: 0.72 ± 0.2 and 0.65 ± 0.07, respectively; P>0.1).
Histological tissue analysis
Inoculated mares were more likely to have placental lesions consistent with placentitis at the time of foaling than control mares (5/6 vs. 0/5; P = 0.02). Histological examination of the fetal membranes and umbilicus revealed placentitis in the region of the cervical star or uterine body in 5 of 6 inoculated mares. One mare had inflammatory lesions in the gravid and nongravid uterine horns and 2 mares had evidence of amnionitis at the time of foaling. There was no histological evidence of disease in the placenta from the inoculated mare that failed to develop ultrasonographic evidence of placentitis. No mare in the control group had evidence of placentitis at the time of foaling.
The results of this study concurred with previously published work [9, 11] in that measurement of CTUP is a useful tool to diagnose ascending placentitis in mares. In contrast, measurement of uterine artery blood flow was not useful to diagnose ascending placentitis in this study. Further, the normal CTUP values for ponies in the current study were similar to those previously reported for light horse mares , indicating that published values can be used as endpoints for diagnosis of clinical and experimentally induced ascending placentitis in pony mares. In contrast, physical examination parameters were not different between mares in either group and uterine artery blood flow was not different between mares in either group. Findings from the study did not support our initial hypothesis as no parameter derived from Doppler ultrasound was different between groups. Although Doppler ultrasound has been used successfully in the prediction of preterm labour in women [20, 21, 36] and values obtained using this procedure were significantly altered at the time of early embryonic loss in the mare , it was unable to detect placentitis in this study. In each of the previous works, relatively small changes were detected early in pregnancy at a time when resistance and pulsatility indices are relatively high and change rapidly as parity increases [16, 23]. In contrast, placentitis is a disease of late pregnancy and both disease onset and diagnosis occur at a time when uterine blood flow is already very high (resulting in a low resistance and pulsatility indices and a high total arterial blood flow) . Thus, in late gestation, any changes in blood flow resulting from disease may be obscured by the overwhelming effects of late gestational demands on the uterine vasculature. It is possible that the utilisation of more sophisticated software for image analysis, as described by Bollwein and others, would enable us to demonstrate differences between groups [13, 29, 37]. However, we aimed to investigate the usefulness of Doppler technology for clinical diagnosis of disease. For this purpose, the use of independent image analysis technology and computer software would be impractical. In a clinical setting, differences between healthy and diseased animals would have to be readily detectable utilising the internal software included in portable ultrasound machines. Furthermore, other studies in mares and stallions have been able to demonstrate differences in blood flow utilising internal software with real-time calculations of resistance and pulsatility indices [15, 16].
The pregnancy outcome of experimentally induced placentitis in the mares in this study was somewhat poorer than previously published , with 2 of 5 infected mares delivering a viable foal in the current study. The study was not designed to specifically investigate treatment efficacy, as no infected, untreated mares were included in the study. However, we attempted to closely approximate all parameters of the previous protocol , including the inoculation method, dose of inoculum, gestational age at inoculation and treatment regimen. One difference between the current study and previous work was the parameters used to initiate treatment. Whereas treatment was previously initiated by either ultrasonographic findings or development of vulvar discharge , treatment in this study was initiated only by an increase in CTUP above published normal values. This parameter was selected in order to compare B-mode and Doppler ultrasonography as diagnostic modalities for ascending placentitis and because changes in CTUP or mammary development are the most common initiating factors for treatment in clinical cases of disease. No mare in either study experienced precocious mammary development prior to onset of other signs. It is possible that these selection criteria may have resulted in a slight delay in treatment and decrease in fetal survivability. Larger studies utilising experimentally induced placentitis in mares or retrospective studies of clinical cases of placentitis are needed to further investigate the effect of treatment onset on pregnancy outcome.
A secondary aim of the current study was to identify a diagnostic tool that could be used to monitor the progress of treated mares with a known infection, and which would enable clinicians to change a treatment approach if the pregnancy was threatened. This was complicated by the small number of mares delivering nonviable foals. However, only gestational length was different between mares delivering viable and nonviable foals. Other measured parameters, including CTUP, fetal heart rate, uterine artery blood flow (resistance index, pulsatility index and total arterial blood flow) and physical examination parameters of the mare were not different in this study, even within 48 h of abortion or delivery.
In conclusion, measurement of CTUP detected experimentally induced placentitis in the current study, whereas measurement of Doppler parameters did not reveal differences between infected and uninfected mares. The CTUP of pony mares in this study was similar to that of light horse mares previously described elsewhere, suggesting that the same parameters should be used for diagnosis of placentitis in ponies and horses.
Authors' declaration of interests
No competing interests exist regarding this work.
Source of funding
This work was funded through an intramural NCSU CVM First Award Grant. Additional funding for student salaries was provided through the Veterinary Scholars Program at North Carolina State University.
The authors would like to thank Jori Vasgaard BS for substantial help during both the data collection and data analysis processes, as well as for her editorial skills.
Bailey – Primary Investigator, grant submission, direct oversight of project, manuscript preparation. Heitzman – Data collection, data analysis, manuscript preparation. Buchanan, Bare, Sper, Archibald – Data collection, mare care, neonatal care. Borst – Histopathology, data analysis, manuscript preparation and editing. Macpherson, Whitacre – intellectual contributions, project design, internal grant review, manuscript preparation and editing.