Twin pregnancy experimental model for transvaginal ultrasound-guided twin reduction in mares

Abstract

Multiple pregnancies are still an important cause of noninfectious abortion, stillbirth, neonatal mortality, and significant delays in reproductive performance in mares. Despite new management techniques, reduction in multiple pregnancies is an ongoing preoccupation and challenge for the equine veterinarian. The aim of the present study was to establish a twin pregnancy experimental model in the mare to study the effectiveness of a transvaginal ultrasound-guided embryonic vesicle injection. Mares in heat were inseminated and then received an embryo at day 7 of the estrous cycle. At days 14 and 30, 53.5% (n = 23) and 23% (n = 10) of the mares, respectively, were carrying twins. Twin pregnancies were reduced at day 30 by transvaginal ultrasound-guided puncture of the embryonic vesicle (control, n = 5) or by transvaginal ultrasound-guided injection (TVUEVI) of 25 mg of amikacin into the embryonic vesicle (n = 5). The TVUEVI treatment had a 40% success rate and no significant variations in progesterone and prostaglandin metabolite were observed. Even though the technique does not seem very effective, the experimental model could be useful for clinical research in embryo reduction and early embryonic loss.

Introduction

Twin pregnancy is undesirable in the equine species because of the serious economic loss it entails. Up to 30% of abortions in brood mares (1–3) are due to multiple pregnancies. Fewer than 9% of pregnant mares carrying twins can carry both fetuses to term (4,5) and with 1 foal surviving after birth in only 21% of cases (6). Not only is the fetal survival rate low, but the long-term future of the offspring is also not favorable. Merkt and Jöchle (7) reported on 104 twin pregnancies with only 16% (n = 17) of the foals reaching the age of 2 y; only 5% of the animals could be trained and none of them could actually race. In addition, twin pregnancy is associated with a higher risk of dystocia and requires more monitoring and assistance compared with single pregnancy.

The difference in size between the 2 embryonic vesicles is an important factor in the outcome of the twin pregnancy. More frequently, twin embryos of dissimilar size locate in the same uterine horn (8), which may result in less nutritional exchange with the mother because of the nonfunctional surface contact between the 2 embryos. This nonfunctional interaction surface usually results in the survival of only 1 embryo by day 40 of pregnancy. Approximately 85% of unilateral twins with a more than 4 mm diameter difference between the 2 embryonic vesicles resulted in the natural elimination of 1 of the twins (9) and normal development of the surviving embryo (10). In bilateral twin pregnancies (in different uterine horns), both embryos survive beyond day 40 of pregnancy more often, but eventually the reduced surface of contact between the chorion and the endometrium results in fetal growth-retarded fetuses, abortion, poor neonatal viability, and stillbirth, in addition to the risk of dystocia associated with twins and reduced fertility in the following breeding season (9). Management of equine twin pregnancies after 40 d gestation is further complicated by the formation of endometrial cups. The endometrial cups secrete equine chorionic gonadotrophin (eCG) (11) that stimulates the formation of secondary corpora lutea, thereby maintaining a high level of circulating ovarian progesterone until day 120 of the pregnancy (12–14). This pregnancy maintenance mechanism means that any medical approach that eventually causes abortion during that period of pregnancy would delay a rapid return to a normal estrous cycle.

Since the introduction of ultrasonography as part of the examination of the reproductive tract in the mare (15), which enables early detection of multiple pregnancies (16), the ability to accurately diagnose and manage twin pregnancies in mares has improved significantly. However, some twin pregnancies are still not diagnosed until after the formation of the endometrial cups, thus jeopardizing the breeding season. Several reasons could explain the failure to correctly detect twin pregnancies: the lack of a breeding history or medical records; an incomplete genital examination; the difficulty in distinguishing between endometrial cysts and young embryonic vesicles with ultrasonography in early pregnancy; examination before day 14 postovulation in mares with asynchronous embryonic vesicles, with 1 vesicle too small to be detected; a poor imaging resolution of the ultrasonographic machine; and, finally, the presence of embryonic vesicles that are closely apposed to each other, making accurate twin pregnancy diagnosis difficult or uncertain on the basis of only 1 examination.

The most effective medical approach to reduce equine twin pregnancy has been manual crush of 1 embryonic vesicle during the mobility phase (< day 16 after ovulation) with a greater than 90% success rate (17–19). In general, the smaller or the most accessible embryonic vesicle is crushed. During the early period of embryonic fixation (> 16 to 30 days postovulation), the success rate of twin reduction is significantly reduced. The size of the uterine horn, and the spatial relationship between the twin embryonic vesicles (unilateral vs bilateral twins) in the uterus influence the reduction success rate. The success (disappearance of 1 vesicle) of manual crushing after unilateral fixation is significantly reduced [25%, (20)], although it is still high in bilateral fixation (20,21). Selective surgical removal of 1 embryonic vesicle has been performed and resulted in the delivery of a single viable foal in 63% (5 out of 8) of mares carrying bicornuate twin concepti (22). Transabdominal ultrasound-guided fetal intracardiac injection of potassium chloride between days 115 and 130 gestation has been assessed with a success rate (single delivered live foal) of 50% (23). Transvaginal ultrasound-guided intravesicular puncture was investigated by several researchers (24–26,12) with variable success (viable conceptus after 10 d), rates: 46%, 31%, 33%, and 20%, respectively. Transvaginal ultrasound-guided aspiration has also been reported (12,24), but the level of technical difficulty required the procedure to be repeated. Recently, Mari et al (27) reported on unilateral twin pregnancy reduction by transvaginal ultrasound-guided aspiration with a success rate of 70% between 16 and 25 d gestation. Reduction of multiple pregnancy is usually safe for the mother, but uterine infection associated with the puncture has been reported (24), suggesting the need for use of an antibiotic. However, others did not observe uterine or abdominal infection after transvaginal-guided puncture (28). Dietary energy restriction has also been proposed to reduce twin pregnancies after day 30 gestation (29). Finally, craniocervical dislocation through a transrectal or flank surgical approach between 60 and 110 d of gestation has been reported (30). However, the success rate of the technique is still to be determined.

Because the prevalence of multiple pregnancies is low and the effective natural mechanism of reduction limited, establishment of a twin pregnancy experimental model would open new opportunities to study clinical approaches to and biological mechanisms of twin reduction in mares. Furthermore, as diagnosis of twin pregnancy is sometime delayed, more information is needed for intervention between days 25 and 40 of the pregnancy. The aim of the present study was to establish a twin pregnancy experimental model in the mare and to compare the effectiveness of transvaginal ultrasound-guided embryonic vesicle injection (TVUEVI) of amikacine (25 mg) with the transvaginal ultrasound-guided embryonic vesicle puncture (TVUEVP) alone.

Materials and methods

Animal

Fifteen vaccinated and dewormed standardbred-cross mares in good condition [550 kg body weight (BW)] and an average age of 9 y were used for the experiment from May 1 to October 1, 2004. Mares were maintained on 16 h of artificial light between December 2003 and April 2004 and on the same breeding management. All procedures were performed in accordance with the Canadian Council of Animal Care recommendations and approved by the University of Montreal Committee of Animal Care and Experimentation.

Experimental design

Groups of 4 or 5 mares were simultaneously induced to estrus, monitored for ovulation, and inseminated. Based on the time of ovulation, 1 mare (recipient) was selected to receive a day 7 embryo collected from another mare of the group (donor). If pregnant on day 28, the recipient was randomly assigned to a treatment (TVUEVI or TVUEVP). If not pregnant, the recipient was resynchronized with other mares of the herd.

Breeding management, embryo transfer, and pregnancy diagnosis

Estrus was induced by injection of prostaglandin (PGF2α, Lutalyse; UpJohn, Orangeville, Ontario), 5 mg/kg BW, IM. Animals were monitored starting on day 4 post-PGF2α injection by daily transrectal palpation and ultrasonography to determine the day of ovulation and the degree of endometrial edema. The number and size of follicles and the size and quality of the corpus luteum (CL) were recorded. When the dominant follicle was ≥ 35 mm in diameter, mares were inseminated with fresh semen from a known fertile stallion and treated with human chorionic gonadotropin (hCG, Chorulon; Intervet, Whitby, Ontario), 2500 IU/kg BW, IV. Ovulation was defined as disappearance of the preovulatory follicles between 2 successive examinations, accompanied by ultrasonographic evidence of a corpus luteum. Semen was collected from the stallion by using a Missouri model artificial vagina (Nasco, Fort Atkinson, Wisconsin, USA) and evaluated for volume, concentration, and progressive motility. The fresh semen was mixed 1:1 (v:v) with skimmed milk glucose-based extender (EZ-Mixin; Animal Reproduction Systems, Chino, California, USA) and each insemination dose contained at least 500 million progressively motile spermatozoa.

Seven days after ovulation, the donors were prepared for embryo collection, according to a standard procedure (31), with 2 L of flushing solution (Vigro; AB Technology, Pullman, Oregon, USA). Recovered embryos were classified (31) and inserted into 0.25-mL straws. Embryos were transferred transcervically into inseminated recipients that had ovulated 7–10 d earlier. The pregnancy was monitored at 16, 21, and 28 d after ovulation by using portable transrectal ultrasonography (ALOKA 900 Alliance Medical, Montreal, Quebec) equipped with 5.0–7.5 MHz transducer.

Data collection

For the 2 experimental groups (TVUEVI and TVUEVP), day 0 being the day of treatment, the pregnancies were monitored by transrectal ultrasonography on days 1, 2, and 5 for heart beat, fetal length, vesicle diameter, echotexture of vesicular liquid, presence of liquid in the uterus, and endometrial edema. Blood samples were taken before treatment (T0) and at 15 min (T15), 60 min (T60), 24 h (T24), 48 h (T48), and 120 h (T120) after treatment to determine the serologic concentrations of progesterone (P4) and 15-keto-13, 14-dihydro-PGF2α metabolites (PGFM).

Progesterone radioimmunoassay

The P4 concentrations were determined by a nonextraction radioimmunoassay (RIAm), using I¹²⁵-P4 (Amersham Biosciences, Oakville, Ontario) as tracer and danazol (sanofi-Withrop, Markham, Ontario) as the steroid displacer (32). The sensitivity of the assay was 0.2 ng/mL. All samples were run in a single assay with an intra-assay coefficient of variation of 9%.

Prostaglandin F2a metabolite assay

The plasma levels of prostaglandin F2a (PGFM) were measured with an enzyme-linked immunosorbent assay (ELISA) (33), using the 13,14-dihydro-15 Prostaglandin F2α EIA Kit (Cayman Chemical; Ann Arbor, Michigan, USA), according to the manufacturer’s instructions. The range of the standard curve for the assay was 7.8 to 1000 pg/mL. The intra-assay coefficient was 4.1% and median inhibition concentration (IC₅₀) was 89 pg/mL.

Puncture and injection of the embryonic vesicle (Day 0)

For the TVUEVI or the TVUEVP (day 28), the mare was placed in a stock with her tail wrapped and the perineum carefully scrubbed. The animal was tranquilized with a combination of xylazine hydrochloride (Anased; Novopharm, Toronto, Ontario), 0.6 mg/kg BW, IV, and acepromazine maleate (Atravet; Ayerst, Montreal, Quebec), 0.03 mg/kg BW, IV. Afterwards, propantheline bromide (Pro-Banthine; Pharmacie Giroux, Saint-Hyacinthe, Quebec), 0.1 mg/kg BW, IV, was used to induce rectal relaxation. The TVUEVI was then performed, using a transrectal ultrasound system (240 Parus Ultrasound Scanner; Pie Medical, Maastricht, The Netherlands) equipped with a 5–7.5 MHz linear transrectal transducer and an external needle guide (Veterinary IVF biopsy attachment; Terumo Medical Corporation, Somerset, New Jersey, USA) equipped with 90-cm silicone tubing plugged to a 20-gauge needle. The ultrasound transducer, covered with a sterile sleeve, was introduced into the fornix of the vagina by pushing the vaginal wall against the uterine horn. The puncture guide on the ultrasound screen was used to select the most effective path for needle placement in the vesicle. The embryonic vesicle was punctured and injected, or only punctured, by maintaining the uterine horn against the transducer while advancing the 20-gauge sterile needle into the embryonic vesicle. Once the needle had been secured in place, 1 mL of sterile saline containing 25 mg of amikacine (Amiglyde-V 250 mg/mL; Wyeth Animal Health, Division of Wyeth Canada, Guelph, Ontario) was injected into the vesicle. The injection of amikacine in the embryonic vesicle was confirmed by visualization of fluid movement within the vesicle while injecting the antibiotic. The 2nd vesicle of the twin pregnancy was visualized on the monitor and left intact.

Statistical analysis and experimental design

Pregnant mares (N = 10) carrying twins were randomly assigned to a treatment group (TVUEVI, n = 5) and a control group (TVUEVP, n = 5). The overall effect of both reducing techniques on P4 and PGFM concentration over time was evaluated by MANOVA with repeated measures for time. Statistical significance was set at P = 0.05.

Results

From 15 mares, 55 uterine flushes were performed at day 7 after ovulation for a recovery rate of 80% (44 embryos). All fresh embryos but 1 (98%) were classified as expanded blastocysts of good quality and transferred. The twin pregnancy rate at days 16 and 28 was 53.5% (23/43) and 43.5% (10/23), respectively. Of the 10 twins on day 28, 8 (80%) were bicornual and 2 (20%) were unicornual pregnancies. The distribution was similar for each group. On day 35, the TVUEVI resulted in 40% (n = 2) single pregnancies, 40 % (n = 2) loss of both embryos, and 20% (n = 1) twin pregnancy. The TVUEVP resulted in a single pregnancy, loss of both embryos, and maintenance of a twin pregnancy in 20% (n = 1), 60% (n = 3), and 20% (n = 1), respectively. Signs of embryonic loss were observed within 24 h after the reduction and were characterized by absence of heartbeat and/or increased echodensity of the embryonic vesicle content without changes in the uterus. The dimensions of the embryonic vesicles are shown in Table 1. There was a mean difference of 5.6 mm and 7.1 mm between the twin embryonic vesicles in the control and treatment groups, respectively, and a mean embryonic vesicle diameter of 35.8 mm and 34.0 mm for the control and treatment groups, respectively.

Table 1.

Characteristics of the embryonic vesicles. C — control group (vesicle puncture only), T — treated group (vesicle puncture and injection of amikacin), V1 — the 1st vesicle and V2: the 2nd vesicle. The diameter of the embryonic vesicles was assessed on the day of the transvaginal ultrasound-guided reduction

Embryonic vesicle diameter before puncture at day 28
Group (C/T)	Mean diameter V1/V2 (mm)	Difference diameter (mm)	Mean diameter (mm)	Bilateral pregnancy
C	34.0/34.5	0.5	34.3	Yes
C	26.0/35.0	9.0	30.5	Yes
C	25.0/25.0	0.0	25.0	Yes
C	19.5/25.5	6.0	22.5	Yes
C	60.5/73.0	12.5	66.8	No
Mean		5.6	35.8
T	24.5/39.5	15.0	32.0	Yes
T	28.5/39.5	11.0	34.0	Yes
T	38.5/40.5	2.0	39.5	Yes
T	29.5/23.5	6.0	26.0	Yes
T	39.0/37.5	1.5	38.3	No
Mean		7.1	34.0

The least square mean levels of P4 for the control and treatment groups were 8.7 ± 1.4, 7.1 ± 1.3, 2.2 ± 2.6, and 2.1 ± 2.6 ng/mL and 8.0 ± 1.2, 8.2 ± 1.1, 8.1 ± 1.3, and 6.4 ± 2.2 ng/mL for days 0, 1, 2, and 5, respectively. In the control and treatment groups, the mean P4 plasma levels were 5.0 ± 1.4 and 7.7 ± 0.9 ng/mL, respectively. However, the statistical analysis showed that there was no significant difference in plasma P4 concentrations between the control and the treated groups (P = 0.2), no change with time (P = 0.1) in both groups, and no interaction between time and treatment (P = 0.2). In the treatment group, in the 2 mares that successfully maintained a single embryo, the plasma P4 concentrations were maintained in 1 case and increased in the other case following the reduction. In the TVUEVP group, the only mare that maintained a single embryo alive sustained a 50% reduction in plasma P4 level (20.0 to 11.0 ng/mL) for the first 2 d and to 9.0 ng/mL on day 5. The least square mean of the PGFM plasma concentrations at time 0, 15, 60 min, 24 h, and 48 h, were 179.8 ± 50.7, 129.3 ± 50.6, 165.7 ± 50.0, 197.4 ± 50.7, and 128.8 ± 98.1 pg/mL for the control group and 50.2 ± 43.0, 136.1 ± 43.7, 154.9 ± 43.6, 55.0 ± 43.0, and 393.8 ± 64.5 pg/mL for the treated group. Analysis of the PGFM data showed that there was no effect of treatment (P = 0.90) or of time (P = 0.21). However, interaction between time and treatment was statistically different (P = 0.02) with the highest value 48 h after treatment. For both P4 and PGFM, there was a significant effect of individual mares (P ≤ 0.05).

Discussion

The embryo recovery of the present study was 80%, which represents the highest expected embryo recovery rate for an embryo transfer program (34). The use of fresh semen of a stallion with high fertility, young mares, and an intensive monitoring program of the donors and recipients could explain the high recovery rate. Following the transfer of embryos, a 53.5% twin pregnancy rate was confirmed by ultrasonography on day 16, which surpassed by far the rate for Thoroughbred and draft mares, which have the highest twin pregnancy rate for the equine species of about 10% between days 14 and 25 (35,36). The mid-size horse breeds (standardbred), like the animals of the present study, have an expected twinning rate of about 5% (day 16). The present experimental model increased the twin pregnancy rate to 10 times that of the natural expected twinning rate and has shown for the first time that it is possible to effectively produce early twin pregnancies in mares. Superovulation could be another method to generate multiple ovulation, multiple embryos at day 7 (37), and potentially, twin pregnancies; however, no data is available at the present time. Different factors (breed, age, and reproduction status) are associated with an increase in twin pregnancy rate (5) and the present model could allow researchers to study different mechanisms for twin reduction in mares.

Of the 23 twin pregnancies at day 16, 65% (15/23) were unilateral and 35% (8/23) were bilateral twin vesicles. Ginther and Griffin (9) reported unilateral fixation of twin embryonic vesicles in about 70% of pregnancies, regardless of the size of the vesicle. When the sizes of the vesicles were taken into account (dissimilar size was defined as a ≥ 4-mm difference between the 2 embryonic vesicles), unilateral fixation in the uterus was more commonly seen (85%) with dissimilar size vesicles (38). After day 16, spontaneous postfixation reduction of unilateral twin pregnancies occurs by day 40 in between 85% and 100% of cases, depending on the size difference between the vesicles (5), compared with no reduction with bilateral fixation for the same period. In the present study, 87% (13/15) of the unilateral twin pregnancies, but none of the bilateral twin pregnancies were spontaneously reduced by day 30. Ginther and Griffin (9) showed that more than half of the postfixation reduction (59%) occurs between days 16 and 20, which is comparable with the results in this study in which more than 50% of twin pregnancies were lost between days 16 and 30. Only 1 of the 4 twin pregnancies with the smallest diameter difference between the twin vesicles (0.0, 0.5, 1.5, and 2.0 mm, respectively) was a unilateral twin pregnancy (Table 1). In the other unilateral twin pregnancy, the difference in diameter between the 2 twin vesicles was 12.5 mm. The proposed experimental model seems to reproduce the physiological situation (5).

Twin pregnancy reduction techniques (manual and transvaginal puncture) before day 20 are effective [> 90%, (19,20,39,40)]. The transvaginal ultrasound-guided intra-vesicular injection of amikacin sulfate for reduction of twin pregnancy was investigated with a view to decreasing the risk of compromising the remaining embryonic vesicle and the mare. The small volume (1.0 mL) was chosen for injection into the embryonic vesicle in order to minimize the increase in size of the injected vesicle and, thus, the disturbance to the other vesicle. Amikacin sulfate, a semi-synthetic aminoglycoside derived from kanamicin, was used as the embryonic toxic substance (pH of 3.5 to 5.5), so that, even with leakage into the uterine lumen, the small volume would be quickly neutralized. Also, the use of an antibiotic would reduce the risk of any uterine infection associated with the puncture (24). Amikacin’s spectrum of activity includes coverage against aerobic gram-negative bacilli and some gram-positive bacteria and amikacin is approved for intrauterine infusion in mares (41).

Although not significantly different than TVUEVP, TVUEVI successfully reduced 40% (2/5) of twin pregnancies. This technique may prove to be a better method for twin reduction but more studies are needed. The success rate was lower than that obtained by transvaginal ultrasound-guided aspiration of the fluid of the embryonic vesicle. Bracer et al (24) succeeded in reducing 3 out 9 unilateral twins and 3 out of 4 bilateral twins between 25 and 45 d gestation. In this case, the pregnancies were assessed 10 d after the procedure. The period between days 20 to 30 is a difficult one for manipulative twin reduction, because at this age, it is very easy to rupture or damage both vesicles (especially unilateral vesicles) during manipulation causing the overall reduction success rate to drop to less than 50% (39). In the present experiment, better results were expected, because 75% (4/5) of the pregnancies of the treated and control group were bilateral twin vesicles. The difficulty of the technique and the operator ability could explain the low success rate. However, increasing the experience of the operator could allow for significant improvement in the results, as it did with other reducing techniques (19). Taking into account the difficulty of the experiment and the small number of mares included in the study, results were comparable with those of previous studies (12,24–26).

Reduction of P4 concentration and uterine contraction could explain the low success rate of the technique. However, no difference in plasma P4 concentration was observed between groups or over time. Pregnancy in the mare is maintained by luteostasis, which is achieved by preventing oxytocin stimulation and prostaglandin secretion by the uterus. Therefore, any uterine traumas would have the potential to stimulate an oxytocin release that could eventually initiate the luteolytic cascade of prostaglandins and cause the loss of pregnancy. Pascoe et al (19) found that twin pregnancy reduction by manual crushing between days 12 and 30 was associated with an increased release of PGFM. However, Veronesi et al (40) showed that reduction of twin pregnancy by embryonic vesicle crushing at day 19 of the pregnancy did not initiate a significant increase of PGFM, even though some variations were observed (30 min after crushing of a fixed embryonic vesicle). Between days 28 and 32 of pregnancy, using a transvaginal puncture technique for twin reduction, the present results showed no significant difference in plasma PGFM concentrations between control and treated mares and no increase at time T15, T60, and T24h, although a significant interaction was measured at time T48. In agreement with the results of Veronesi et al (40) with manual crushing reduction, the present results highlight the lack of luteolytic effect of the transvaginal puncture to induce PGF2α and, eventually, pregnancy loss.

The present experimental model was effective in producing twin pregnancies in standardbred-cross mares and allowed us to study the effectiveness of transvaginal ultrasound-guided embryonic vesicle injection (TVUEVI) in the reduction of twin pregnancies. The high efficiency of this model to produce twins provides the opportunity to study basic physiologic mechanisms of natural multiple pregnancy reduction, embryonic mobility in the uterus, and reducing techniques.

Early detection by transrectal ultrasonography and manual crushing of 1 of the embryonic vesicle are still the key elements in successful management of twin pregnancy. In cases of unilateral twin pregnancy with embryonic vesicles of similar size diagnosed between days 20 to 30 gestation, the use of TVUEVI might be a medical option, but more research and clinical data are needed before fully recommending it.