Dr Jacobson's present address is: Section of Reproductive Studies, New Bolton Center, School of Veterinary Medicine, University of Pennsylvania, Kennett Square, Pennsylvania 19348, USA. Dr Hayden's present address is: Department of Veterinary Clinical Sciences, Center for Veterinary Health Sciences, Oklahoma State University, Stillwater, Oklahoma 74078, USA.
Reasons for performing study: Transvaginal ultrasound-guided follicle aspiration (TVA) is performed clinically but there is little information available on complications associated with this procedure.
Objectives: It is possible that TVA is associated with damage to the ovary and may induce peritonitis or peritoneal adhesions. This study was conducted to determine the effect of repeated TVA on mare health and ovarian status.
Methods: Thirty-two mares were used for oocyte recovery via repeated TVA over a 3 year period; different mares were used each year. In Year 1, ovarian status was monitored in 11 mares by transrectal palpation and ultrasonography. In Year 2, 6 of 11 mares underwent abdominocentesis and were examined by laparoscopy after one TVA and again after multiple TVAs. In Year 3, 10 mares underwent multiple TVAs with either a 15 or a 12 gauge needle and the ovaries were removed for examination.
Results: Four hundred and twenty-seven aspiration sessions (390 via TVA and 37 via needle placement through the flank) and 3202 follicle punctures (3161 TVA and 41 flank) were performed. One mare developed an ovarian abscess. Transient rectal bleeding was evident after 16% of TVA sessions. No adhesions were found on laparoscopic or gross examination of ovaries and there were minimal changes on histological evaluation.
Conclusions: Follicle aspiration carries a small possibility (<0.5%) of ovarian abscess formation. There is a possibility of rectal abrasion or puncture but little gross or histological damage to the ovary.
Potential relevance: These results provide a basis for using prophylactic administration of antibiotics after TVA and for advising mare owners of the rare but potential complications associated with the procedure.
Transvaginal ultrasound-guided follicle aspiration (TVA) has been performed in mares for research purposes for over 20 years  and is now being used clinically to obtain oocytes for intracytoplasmic sperm injection . Oocyte recovery may be performed not only in subfertile mares but also in fertile mares being bred to stallions with limited semen quality or availability . Therefore, it is important to determine the effect of repeated TVA on mare health and ovarian status.
Aspiration of the stimulated dominant preovulatory follicle may be performed to provide access to one, or possibly 2, mature oocytes per cycle. However, this technique was associated with lower blastocyst production after intracytoplasmic sperm injection than occurred when immature oocytes were collected from all visible follicles on the ovaries . This latter procedure requires multiple needle punctures through the ovarian stroma and thus carries the potential for ovarian damage and subsequent formation of adhesions. The ultrasound probe is introduced into the vagina through the vestibule, which may be heavily contaminated with bacteria . Because the aspiration needle is then passed repeatedly from the probe through the vaginal wall and into the peritoneum, TVA would seem to also carry the potential for inducing peritonitis and ovarian infection.
Little information is available on the effect of repeated TVA in mares. Normal pregnancy rates were reported in mares after aspiration of small numbers (1–25) of dominant or small follicles via TVA [6–8]. There is one previous report of death of a mare due to haemorrhage resulting from arterial puncture during TVA . Bogh et al.  reported the development of an ovarian abscess in one of 4 pony mares subjected to repeated TVA (average of 121 follicles aspirated per mare) over an 8-year period. To the best of our knowledge, the latter is the only report available on the effect of repeated TVA of immature follicles on ovarian status in mares.
The present studies were performed to evaluate the health and ovarian status of mares subjected to repeated TVAs for collection of oocytes for our research programme at Texas A&M University. Results of oocyte recovery, in vitro maturation and blastocyst production after intracytoplasmic sperm injection have previously been reported for a portion of the aspiration sessions included in Year 1 .
Materials and methods
Year 1. Assessment of ovarian changes via palpation and ultrasonography per rectum
Mares: This was a retrospective study of aspirations performed from March to December 2008. Eleven Quarter Horse-type mares, 5–16 years of age and weighing 400–500 kg, were used. The mares were housed outside in paddocks and fed hay ad libitum with free choice water. All experimental procedures were performed according to the United States Government Principles for the Utilization and Care of Vertebrate Animals Used in Testing, Research and Training and were approved by the Laboratory Animal Care Committee at Texas A&M University.
Follicle aspiration: Initially, 7 mares were subjected to TVA sporadically as the procedure was developed. Starting in May, all mares had TVA of immature follicles once every 6–44 days for the remainder of the year, most frequently with a 14 day inter-aspiration interval. One preovulatory follicle was aspirated through the flank. Palpation and ultrasonography per rectum were performed weekly.
Transvaginal follicle aspiration was performed as previously described . Detailed procedural methods are provided in the Supporting Information. An Aloka 500 ultrasound machine with a 5-mHz sector scan ultrasound probe was useda. A 12-gauge double lumen oocyte aspiration needleb was used for the majority of aspirations. However, on 11 occasions a 15- or 16- gauge double lumen needlec was used, as this was being evaluated to determine whether the smaller gauge might pass more easily through the ovarian stroma and thus increase the ease of the procedure. When possible, each follicle was flushed 6–8 times and the follicle evacuated by rotating the needle after each flush, while the ovary was moved gently per rectum.
If bleeding from the rectum was noted, the aspiration was stopped and the rectum examined by palpation and speculum examination. If bleeding continued or a tear was suspected, 2 l of mineral oil was administered by nasogastric intubation. Procaine penicillin Gd (20,000 iu/kg bwt, i.m.) was administered twice daily and gentamicine (6.6 mg/kg bwt, i.v.) once daily for a maximum of 3 days. The mare was kept in a stall with no food overnight and evaluation of the rectum was repeated the next morning.
The oocyte recovery rate was calculated based on the number of oocytes recovered per number of follicles punctured. Every follicle visibly punctured by the aspiration needle was counted, regardless of the completion of adequate evacuation and flush. Oocytes were recovered from aspirates and matured in vitro as previously described .
For aspiration of the preovulatory follicle, mares having a follicle ≥29 mm in diameter were treated with human chorionic gonadotrophinf (1500 iu/kg bwt i.v.) and/or biorelease desloreling (1.5 mg/kg bwt i.m.). Aspiration was performed 24–35 h later via a cannula placed through the flank as previously described .
Year 2. Assessment via abdominocentesis and laparoscopy
Mares: This was a prospective study performed from February to October, 2009. Eleven Quarter Horse-type mares 6–14 years of age, separate from the mares used in Year 1, were maintained under the same conditions.
Follicle aspiration: Mares underwent repeated TVA of all immature follicles, with a minimum of 9 days between aspirations. A 12 gauge needleb,c,h was used for the aspirations, except on 8 occasions on which a 15 gauge needle was used when a 12 gauge needle was not available. Evaluation of mares for ovarian abnormalities, management of bleeding from the rectum, oocyte maturation in vitro and occasional preovulatory follicle aspiration through the flank were performed as for Year 1.
Abdominocentesis and laparoscopy: Six mares were randomly assigned to undergo abdominal laparoscopy, on either Day 3 or Day 10 after their first TVA. Laparoscopy was not performed on both days in each mare as the effects of the previous laparoscopy may have confounded interpretation of changes due to the TVA. Abdominocentesis was performed on Day 3 in all mares, as described in the Supporting Information, and was repeated on Day 10 in those mares undergoing the Day 10 laparoscopy. The 6 mares underwent a second exploratory laparoscopy at the end of the year, a minimum of 40 days after the final TVA session.
Abdominal laparoscopy was performed through each flank to visualise the ipsilateral ovary, as described in the Supporting Information. Still photos or video recordings were used to document findings. In some mares, transrectal palpation was used to manipulate the ovary to fully evaluate all surfaces via laparoscopy.
Embryo recovery after insemination: The mare (TR) showing the most ovarian changes on laparoscopy was inseminated the following breeding season. Uterine flush for embryo recovery was performed 7 days after ovulation on 3 cycles; for the subsequent 2 cycles, the mare was inseminated and examined for pregnancy by ultrasonography per rectum starting 11 days after ovulation.
Data analysis: Correlation between peritoneal fluid analysis values and number of follicles punctured was analysed by Pearson's product moment correlation using SigmaPlot statistical software.
Year 3. Gross and histological assessment of excised ovaries after follicle aspiration with 12 or 15 gauge needles
Mares: This was a prospective study performed from March to September 2010. Ten Quarter Horse-type mares, 4–12 years of age and weighing 400–500 kg, were used. These mares were separate from the mares used in Years 1 and 2 and were maintained similarly.
Follicle aspiration: Mares were randomly assigned to either a 15 gauge needle group (15G) or a 12 gauge needle group (12G), with 5 mares per group. The mares remained in the same group throughout the breeding season. Mares were subjected to TVA with the corresponding-size needle every 10–24 days. Preovulatory follicles were occasionally aspirated through the flank.
Embryo recovery after insemination: Six mares were inseminated for embryo recovery between aspiration sessions to obtain embryos for another project. In 5 mares, a follicle ≥30 mm diameter was seen on ultrasonography per rectum one week after TVA (n = 2) or on the day of TVA (and this follicle was left unaspirated; n = 3). If no uterine oedema was visible, the mare was treated with prostaglandin F2α, 5 mg i.m. Mares were treated with hCG, 2000 iu i.v. (n = 3) or biorelease deslorelin, 1.5 mg i.m. (n = 2), on the day the follicle was first seen or up to 2 days afterwards, depending upon the availability of semen and the desired day of embryo flush. Mares were inseminated within one day of gonadotrophic treatment and examined daily by ultrasonography per rectum until ovulation was observed. The sixth mare had a 40 mm follicle at the of time of TVA that was left unaspirated; when examined 3 h after TVA, the follicle had collapsed and the mare was inseminated at that time. In 5 mares, uterine flush for embryo recovery was performed 7 days after ovulation. In one mare, pregnancy examination was performed by ultrasonography per rectum 13 days after ovulation.
Oocyte recovery from excised ovaries: At the end of the year, 8 mares were ovariectomised via colpotomy as described in the Supporting Information and 2 mares were subjected to euthanasia with pentobarbitalf (100 mg/kg bwt i.v.) and the ovaries recovered post mortem.
Recovered ovaries were examined grossly and dissected for oocyte recovery. Sections of ovary including any apparent anomalous areas were processed for histopathology.
Oocyte collection was performed as previously described by scraping all visible follicles . Recovered oocyte–cumulus complexes were classified as compact (Cp), expanded (Ex) or degenerating depending on the expansion of both mural granulosa and cumulus as described previously . Oocytes classified as Ex or Cp were matured in vitro and assessed for maturation to metaphase II.
Light microscopic examination of ovarian tissue was performed on formalin-fixed, paraffin-embedded, haematoxylin and eosin-stained tissue sections.
Data analysis: The difference in oocyte recovery rate for the 2 needle size groups was determined by Chi-square analysis. The number of follicles processed for oocyte recovery in excised ovaries between groups was determined by the Mann–Whitney rank sum test.
One flank aspiration and 153 TVA sessions were performed (range 11–17 TVA sessions per mare). An average of 8.7 follicles were punctured per mare per TVA session (range 2–26). A total of 1337 follicles were aspirated, for an average of 122 (range 62–224) follicles aspirated per mare. The oocyte recovery rate was 671/1337 (50%). The percentage of oocytes that matured to metaphase II in vitro was 58% (375/651). The 15-gauge needle slightly facilitated ovarian puncture but greatly slowed fluid transport; thus the 12-gauge needle was used preferentially for subsequent TVA sessions.
Bleeding from the rectum was recorded during 25 of the 153 sessions (16%); on 3 of these occasions, the needle appeared to puncture the rectum during TVA. Bleeding was minor and transient except in one mare, which was stalled and treated with mineral oil and antibiotics as described above. No mare showed abnormalities (depression, inappetence, colic, etc.) following an incident of rectal bleeding. The mare that was stalled had normal temperature, pulse and respiration during this period. The ovaries of one mare became less freely moveable over time. This mare bled from the rectum on 4/15 aspiration sessions (27%), suggesting that difficulty in ovarian manipulation may have contributed to rectal bleeding. The remaining mares remained healthy with no apparent ovarian abnormalities.
Thirty-one flank and 131 TVA sessions were performed (range 5–17 TVA sessions per mare). An average of 8.3 follicles were punctured per mare per TVA session (range 1–22). A total of 1091 follicles were aspirated for an average of 99 per mare (range 27–171). The oocyte recovery rate was 40% (434/1091); 70% (292/420) of cultured oocytes matured to metaphase II in vitro.
Bleeding from the rectum was recorded during 29 TVA sessions (22%), 8 of which occurred in one mare. On 4 occasions the needle was felt to have gone through the rectum. Bleeding was minor and transient except in one case; this mare was treated with mineral oil and one day of antibiotic administration. No mare showed abnormalities (depression, inappetence, colic, etc) following an incident of rectal bleeding and the mare that was stalled had normal temperature, pulse and respiration during this period.
One mare's ovaries appeared to be less moveable at the end of the year. This mare had a scar on the flank, indicative of a prior flank approach surgery, and had bleeding from the rectum on 4 occasions in 17 aspiration sessions (24%).
Abdominocentesis: The results of abdominocentesis are given in Table 1. Three days after the mares' first TVA, all samples had high RBC counts and in 2 of the 6 mares WBC counts indicated mild inflammation (>10,000 cells/µl). There was no correlation between the numbers of follicles aspirated and WBC or RBC concentrations (P>0.1).
Table 1. Results of abdominocentesis 3 and 10 days after the first TVA session
No. fols asp
Total protein (g/l)
*All 6 mares underwent abdominocentesis 3 days after the first TVA session. †n/a = Not analysed. ‡Only the 3 mares scheduled for laparoscopy on Day 10 underwent abdominocentesis on Day 10. -; = Unable to obtain fluid for analysis; No. fols asp = number of follicles aspirated.
Fluid was obtained from only 2 of 3 mares on Day 10. The only abnormality was a slightly elevated RBC count and low fibrinogen concentration in one mare.
Abdominal exploratory laparoscopy: On Day 3, needle puncture sites were noted on the ovaries via laparoscopy in 2 of 3 mares and there was a haemorrhagic area in the mesovarium of the left ovary in one mare (Fig 1).
On Day 10 the right ovary of one mare had a diffusely roughened surface (Fig 2) and the left ovary possessed one visible needle puncture site; all other findings were normal.
Laparoscopy was repeated on the above 6 mares 244–274 days after the first TVA and 40–75 days after the final TVA session. The mares had undergone 11–17 TVA sessions (76–171 punctures per ovary) and 1–6 flank aspirations during the year. In 4 mares there was apparent increased opacity of the mesovarium over one or both ovaries, so that the normal ribbing of vessels on the surface of the ovary was not visible (Fig. 3). Mare CH had apparent roughening of the surface of the right ovary. There was no evidence of adhesion formation in any mare.
The mare with the most vivid apparent thickening of the mesovarium (Fig 3), Mare TR, was inseminated on 5 cycles the following breeding season. Four embryos were recovered on uterine flush on the first 3 cycles and the mare was pregnant on the fifth cycle. At least one pregnancy resulted from ovulation from each ovary.
Five flank aspirations and 106 TVA sessions were performed (range 5–12 TVA sessions per mare). An average of 6.9 follicles were punctured per mare per TVA session (range 1–18). A total of 733 follicles were aspirated for an average of 73 follicles per mare (range 38–99). The oocyte recovery rates were 38% (127/337) and 48% (189/396) for the 15G and 12G needle groups, respectively; this difference was significant (P<0.01). The overall oocyte maturation rate was 70% (124/176); this was 73% (58/80) for 15G and 69% (66/96) for 12G (P>0.1). Bleeding from the rectum was recorded on 10 occasions (9%), 2 associated with apparent needle puncture of the rectum, and was transient in all cases. On one occasion, bruising of the vestibule floor was noted. With the exception of Mare PS, detailed below, no abnormalities were noted on ultrasonography or palpation per rectum.
One mare (LS; 15G group; 10 TVA sessions, 80 follicles aspirated) was subjected to euthanasia for ovary removal due to a severe cut on her forehead suffered at pasture.
One mare, Mare PS (15G group), was donated because of ataxia and unthriftiness related to a previous head injury. The right ovary was difficult to manipulate per rectum. A mottled echogenic structure approximately 35 mm in diameter was seen on this ovary 14 days after the first aspiration. The structure was still present 14 days later and prostaglandin F2α (10 mg/kg bwt i.m.) was administered. Eight days afterward, the structure was 45 mm in diameter and the contents sampled via flank aspiration. The recovered fluid was brown and contained numerous RBC but few WBC. There was no bacterial growth on culture but it was noted that the gentamicin in the heparinised medium used during aspiration may have interfered with bacterial growth. Eight days later, one densely echogenic structure and one irregularly echogenic structure were observed on the right ovary (Fig 4a). Follicle aspirations were discontinued in the mare owing to the possibility of ovarian abscess. There were no systemic signs of illness. Because of her unthriftiness and ataxia, the decision was made not to perform ovariectomy via colpotomy and the mare was subjected to euthanasia to remove the ovaries for examination. Euthanasia was performed 66 days after the initial aspiration (52 days after the echogenic structure was first detected). The left ovary was normal. Sectioning of the right ovary exposed an abscess 5 cm in diameter with a fibrous wall and a 3 × 2 cm caseous centre (Fig 4b), from which Streptococcus zooepidemicus was isolated, and an adjacent 3 cm focus of clotted blood with minimal inflammation.
Examination of excised ovaries: As outlined above, ovaries from Mares LS and PS were removed after euthanasia 11 and 22 days after their last TVA session, respectively. Ovaries from all other mares were removed via colpotomy 13–29 days (average 17.9 days) after their last TVA session. No adhesions were palpable on examination via colpotomy. Evaluation of the right ovary from Mare PS is described above. For all other ovaries, gross evaluation revealed presumed needle puncture sites on the surface, more prominent on ovaries excised via colpotomy (Fig 5), probably due to the increased pressure within the ovary when the ecraseur is applied, causing bleeding into the puncture sites. The left ovary of mare SK (12G group) had a roughened area on the tunica albuginea.
The average number of follicles sectioned per pair of ovaries was 14.2 for the 15G group and 11.8 for the 12G group (P>0.1). There was no significant difference in oocyte recovery rates between the 15G and the 12G groups (59/71, 83% and 35/47, 74%, respectively; P>0.1). A total of 79 Ex, 10 Cp and 5 degenerating oocytes were recovered; there was no significant difference in oocyte classification between groups. In vitro maturation rates were 61% (46/75) for Ex and 0% for Cp.
Histopathology of ovarian sections revealed evidence of past haemorrhage with a greater number of siderocytes in the 15G group than in the 12G group. Collapsed follicles lined with atypical luteal tissue were observed in both groups. Other collapsed follicles without lining were associated with red blood cells, siderocytes and fibrotic tissue (Fig 6). Additional observations considered to be incidental were evidence of fresh haemorrhage in the cortex in ovaries removed via colpotomy, nodules of adrenal cortical tissue in 2 mares and one eosinophilic, presumed parasitic, granuloma.
Embryo recovery: Four embryos were recovered from 5 mares on uterine flush 7 days after ovulation. The mare from which no embryo was recovered was the mare that was inseminated after follicle collapse on the day of TVA. Mares that produced embryos had undergone 1, 2, 5 and 5 TVA sessions at the time of insemination. The sixth mare (5 TVA sessions) was pregnant on ultrasonography per rectum on Day 13 after ovulation.
In these studies, 390 TVA and 37 flank follicle aspiration sessions were conducted on 32 mares, for a total of 3161 follicles punctured by TVA and 41 follicles punctured by flank aspiration. The TVAs were performed by 7 different operators over the 3 years of the study. The most notable lesion was an ovarian abscess in one mare (Mare PS). The abscess was found after aspiration of a haemorrhagic structure that formed after a TVA, in an ovary noted on initial examination as difficult to manipulate. Nevertheless, this finding, combined with the similar report of Bogh et al. , indicates that abscess formation is a rare but potential consequence of follicle aspiration.
Bleeding from the rectum was also observed; this was slight and transient in all but 2 cases and those mares recovered without incident. However, the possibility of rectal tear exists whenever transrectal palpation is conducted and remains a potential complication of TVA. In 9 cases the rectum was punctured with the needle during TVA and these all resolved without incident.
With the exception of Mare PS, all mares continued to have normal follicle growth and, if follicles were not aspirated, normal follicle ovulation occurred. Repeated TVA did not appear to affect oocyte development, as in Year 3 the number of follicles per ovary, oocyte recovery rate, oocyte classification and rate of maturation to metaphase II in vitro from ovaries recovered after multiple TVA sessions were consistent with previous findings from slaughterhouse-derived ovaries [13–15]. Pregnancy rates were normal in cycles in which mares were inseminated (9 pregnancies/11 cycles). However, only 7 of the 32 mares used in the studies were inseminated for pregnancy evaluation, thus further work is needed to accurately determine whether any loss of fertility is associated with repeated TVA of all visible follicles.
Peritoneal fluid analysis indicated that 2 of 6 mares had mild peritoneal inflammation after TVA. This may occur after any procedure involving entry into the peritoneum . While this was only a small sampling of the mares, the peritoneal fluid changes appeared to resolve by 10 days after the procedure in the 3 mares examined.
In 2 mares, adhesions of the ovary to the body wall were suspected when the ovary became difficult to manipulate on palpation per rectum. This was confounded by the fact that the mares were obtained from dealers and may have had previous procedures performed on their reproductive tracts. One of these mares (Year 2) had a scar from a previous flank laparotomy; the other (Year 1) was not examined for a scar. In the third year, in which mares were obtained from private sources, no changes in ovarian mobility were noted.
When ovaries were evaluated either by laparoscopy (Year 2) or by excision and evaluation (Year 3), there were few lesions observed. No adhesions were seen between the ovaries and adjacent structures. The apparent mesovarian thickening seen on laparoscopy at the end of Year 2 may relate to organisation of previous haemorrhage in the mesovarium, as seen in one mare in the first laparoscopy, or to seasonal changes. None of the 18 ovaries examined after removal at the end of Year 3 had opacity of the mesovarium; however, ovaries were excised in August and September in Year 3, whereas the laparoscopies were performed in November and December in Year 2.
On histological examination, with the exception of the ovaries from Mare PS, there was no evidence of oophoritis or ovarian inflammation. The atypical luteal tissue noted on histological examination probably represented luteinisation of follicle walls after aspiration; increase in progesterone after aspiration of immature follicles, indicating luteinisation, has been noted previously [4,17]. The increased numbers of siderocytes in the 15G group may reflect the different tip (etched for visualisation by ultrasound, and thus with an irregular surface) on this needle. This, combined with the significant difference in recovery rate, suggests that the 12 gauge needle was preferable for TVA.
In conclusion, clinical, gross and histological findings indicate that repeated TVA sessions do not cause peritonitis, adhesions of the ovary or degeneration or inflammation of the ovarian tissue and have a <0.5% incidence of ovarian infection (one case of abscess formation in 390 aspiration sessions; >3000 follicles aspirated). The potential for infection might be reduced in clinical cases by prophylactic administration of antibiotics. Follicle growth and oocyte viability appear to be unaffected by repeated aspiration of all visible immature follicles on the ovary. The results of these studies provide a basis for advising mare owners of the rare but potential complications associated with performing follicle aspiration for oocyte recovery.
Authors declaration of interests
No competing interests have been declared.
Source of funding
This work was supported by the Link Equine Research Endowment Fund, Texas A&M University, and Ms Kit Knotts.
I.C.V., C.A., C.C.J. and K.H. contributed to study design, data collection, study execution, data analysis and interpretation. J.D.H., Y.H.C. and S.S.H. contributed to data collection and study execution. J.F.E. contributed to data analysis and interpretation.
a Boland Vet Sales, Keller, Texas, USA.
b Cook Veterinary Products, New Buffalo, Michigan, USA.
c Casmed International Limited, Epsom, Surrey, England.
d Durvet, Blue Spring, Missouri, USA.
e Vedco Inc, St. Joseph, Missouri, USA.
f Schering-Plough Animal Health, Union, New Jersey, USA.
g BETPharm, Lexington, Kentucky, USA.
h Mila International Inc, Erlanger, Kentucky, USA.