Theriogenology 59 (2003) 1181±1187

Association of uterine edema with follicle waves

around the onset of the breeding season
in pony mares
E.D. Watsona,*, R. Thomassena,b, E. Nikolakopoulosa,c
a
Department of Veterinary Clinical Studies, Easter Bush Veterinary Centre, University of
Edinburgh, Easter Bush, Roslin, Midlothian EH25 9RG, UK
b
Department of Reproduction and Forensic Medicine, The Norwegian School of
Veterinary Science, Postbox 8146 Dep, 0033 Oslo, Norway
c
44 Ki®ssias Avenue, Athens 115 26, Greece
Received 19 December 2001; accepted 1 May 2002

Abstract

During spring transition, when estrus may be exhibited for prolonged periods, it is important for
veterinarians and stud farm personnel to be able to predict whether a large follicle will ovulate or
regress. It is thought that the presence of ultrasonically detectable uterine edema indicates that a
follicle will ovulate, however, there is little evidence to support this. In the present study, 16 mares
were regularly examined by transrectal ultrasonography to follow growth and regression of follicles
from seasonal anestrus in February until second ovulation. Blood samples were collected daily for
measurement of estradiol concentrations when a large ovarian follicle was present. Estrous-like
uterine edema was detected during 7 of 11 (64%) anovulatory follicle waves, in 12 of 14 (86%) mares
before their ®rst ovulation, and in 100% of mares before their second ovulation. Uterine edema was
®rst detected 43  6:7 days before ®rst ovulation. Large anovulatory follicles tended to be present for
longer periods of time than ovulatory follicles. Uterine edema was present for a signi®cantly greater
proportion of time in the presence of a large follicle at second ovulation than at ®rst ovulation
(P < 0:05) or for anovulatory follicles (P < 0:01). Peak plasma estradiol concentrations and mean
plasma estradiol concentrations were signi®cantly higher (P < 0:001) when a dominant preovulatory
follicle was present compared with a dominant anovulatory follicle, but there was no difference in
estradiol concentrations between ®rst and second ovulations. It was apparent, therefore, that uterine
edema was not a reliable indicator of follicular steroidogenic competence, or of whether the follicle
would ovulate.
# 2002 Elsevier Science Inc. All rights reserved.

Keywords: Mare; Uterine edema; Follicles

*
Corresponding author. Tel.: ‡44-131-650-6236; fax: ‡44-131-6588.
E-mail address: elaine.watson@ed.ac.uk (E.D. Watson).

0093-691X/02/$ ± see front matter # 2002 Elsevier Science Inc. All rights reserved.
PII: S 0 0 9 3 - 6 9 1 X ( 0 2 ) 0 1 1 6 1 - 5
1182 E.D. Watson et al. / Theriogenology 59 (2003) 1181±1187

1. Introduction

In winter, most mares are anestrus and exhibit small hard ovaries with little follicular
activity. In the transitional period between winter anestrus and normal cyclicity, mares
often go through a period known as ``shallow anestrus'' or ``spring transition''. It is
reported that during this time large follicles grow on the ovaries in approximately 50% of
mares, and often reach preovulatory size, but fail to ovulate [1]. An average of 3±4 waves of
single large (<30 mm) follicles grow at intervals of approximately 10 days before one of
these follicles ®nally ovulates [1,2]. In the remaining mares the follicles fail to reach
35 mm in diameter and do not show a clear wave pattern. In spring transition, mares may
remain in estrus for protracted periods of time. However, their circulating concentrations of
estradiol are low, because these large anovulatory follicles are steroidogenically incom-
petent [3,4]. At this time, when the mares are showing estrus, it is dif®cult to deduce from
transrectal ultrasound examination whether the large ovarian follicle will ovulate or
regress. This can result in mares being mated repeatedly over a protracted period before
the ®rst ovulation in spring.
There is clearly a need for an accurate predictor of the ovulatory potential of a follicle
prior to the onset of regular cyclicity. Concentrations of estrogen conjugates in urine have
been shown to increase as ®rst ovulation approaches [5], but these measurements are not
readily available to veterinary practitioners or stud farms. In another study, mares appeared
not to show estrous-like uterine edema more than 15 days before ®rst ovulation in spring [6].
Administration of estradiol to ovariectomized mares induces uterine edema [7], and it has
been suggested that the presence of uterine edema is a reliable indicator of the estrogenic
competence of the dominant follicle [8]. This, therefore, led to the suggestion that the
presence of uterine edema can be used as an aid in predicting whether a large follicle will
ovulate [8]. However, there have been no studies that record ovarian and uterine ®ndings in
detail in spring transition and correlate these with circulating estradiol concentrations.
In the present study, changes in uterine echogenicity were recorded throughout the
spring transition period and in the early breeding season in a group of pony mares and
compared with circulating concentrations of estradiol. The objective of the study was to
critically evaluate if uterine edema could be used as a reliable indicator of whether a large
dominant follicle would ovulate.

2. Materials and methods

Sixteen pony mares of mixed breeding, weighing 240±450 kg and aged 3±20 years were
used. The mares' ovaries and uteri were examined by transrectal ultrasonography (5 mHz
transducer) on every second or third day from the beginning of February until ®rst
ovulation. Eleven of the mares were also studied at their second ovulatory estrus. Follicle
diameters and locations were recorded in an ovarian sketch. When the follicles were not
spherical, the width and height were recorded, and the average of these was used as the
diameter of the follicle. After a follicle of 25 mm in diameter was detected, the mares were
examined daily. Uterine echogenicity was scored midway between the bifurcation and tip
of the uterine horn as (1) uterine edema similar to that seen in estrus, where endometrial
 E.D. Watson et al. / Theriogenology 59 (2003) 1181±1187 1183

folds were apparent, with anechoic areas visible within the tissue, (2) slight uterine edema
where there was some heterogeneity of the tissue, but no obvious endometrial folds, or (3)
no uterine edema, where the uterus appeared homogeneous with no anechoic areas. All of
the authors were involved in ultrasonographic scanning of the mares, but one of the authors
(E.D.W.) was present at the majority of the examinations and checked for consistency
among operators. However in the results, ``slight uterine edema'' was included with ``no
uterine edema'', as it was likely that there were greatest individual operator differences in
this category. Blood samples (20 ml) were collected from the jugular vein into evacuated
heparinized tubes on the same days as the ovarian examinations. Plasma was separated by
centrifugation at 2000  g for 15 min at 4 8C and stored at 20 8C until assayed for
estradiol concentrations. This sampling regime was maintained until the second ovulation
of the breeding season.

2.1. Estradiol assay

Concentrations of estradiol were measured in an extraction radioimmunoassay using a

previously validated technique [9]. The limit of detection of the assay was 2 pg/ml and the
intra- and interassay coef®cients of variation were 4.6 and 7.8%, respectively. Average
recovery of estradiol from plasma was 91%.

2.2. Statistical analysis

Three different types of follicle waves were included: an anovulatory wave, the ®rst
ovulatory wave, and second ovulatory wave. The anovulatory wave was chosen at random
and was the ®rst, second, or third anovulatory wave. For statistical analysis, ``slight uterine
edema'' was pooled with ``no uterine edema''. The effect of wave type on the number of
days that uterine edema was present, the number of days that the large follicle (>25 mm)
was present on the ovaries, and on estradiol concentrations (peak concentration measured in
daily samples during the growth phase (>25 mm) of a follicle wave, and mean daily
concentrations over this period), and the proportion of time that uterine edema was observed
in the presence of a large follicle (number of days of uterine edema/number of days when
large follicle present) were analyzed using a one-way ANOVA with least signi®cant
difference mean comparisons, where appropriate. The follicle diameter of 25 mm was
chosen as it has been reported that this is the size of the dominant follicle when pony mares
come into estrus [10]. Mean concentrations of estradiol were calculated for each mare that
showed uterine edema (a) on the days that she had uterine edema and (b) on the three
consecutive days before and after detection of edema. Differences in concentrations were
detected using a paired t-test. Correlations between peak and mean estradiol concentrations
were analyzed by Pearson's correlation test. Results are reported as mean  S:E:M.

3. Results

Anovulatory follicular waves, where a single dominant follicle reached at least 30 mm

with a growing, static, and regressing phase, were detected in 11 (69%) of the 16 mares.
1184 E.D. Watson et al. / Theriogenology 59 (2003) 1181±1187

Table 1
Detection of follicular wave (>25 mm), uterine edema, and plasma estradiol concentrations in mares

Type of follicle wave

Anovulatory First Second

ovulation ovulation
Mares (n) 11 14 11
Days follicle wave persisted >25 mm (mean  S.E.M.) 16.3  2.5 12.4  0.9 12.7  1.3
Days uterine edema detected in the presence of a 2.8  0.9 3.1  0.6 4.5  0.6
follicle >25 mm (mean  S.E.M.)
Proportion of time uterine edema detected in the presence 17a 25b 35c
of a follicle >25 mm (%)
Peak estradiol concentrations (pg/ml) in the presence 9.0  2.9d 30.1  2.5e 29.4  3.3e
of a follicular wave (mean  S.E.M.)
Mean estradiol concentrations (pg/ml) in the presence 5.9  1.4f 17.1  1.4g 16.6  1.5g
of a follicular wave (mean  S.E.M.)
Within a row, different superscripts represent a signi®cant difference: a, b (P < 0:05); a, c (P < 0:01); b, c
(P < 0:05); d, e (P < 0:001); and f, g (P < 0:001).

The 11 mares had 1±7 anovulatory follicular waves prior to the ®rst ovulatory follicle, with
a mean of 3:6  0:6 anovulatory waves. In the remaining ®ve mares no follicles grew
greater than 30 mm in diameter and there was no clear wave pattern before growth of the
preovulatory follicle. The median date of ®rst ovulation in the 16 mares was May 3, and the
dates ranged from April 9 to May 12. Data were available from 11 anovulatory follicle
waves, 14 ®rst ovulations, and 11 second ovulations. Waves were not included in the data
analysis if the mare had no anovulatory waves, if more than one large follicle was present
during a wave, or if data collection was incomplete.
Anovulatory follicle waves (presence of a follicle >25 mm) lasted an average of 16.3
days (range 5±27 days; Table 1). Uterine edema (Fig. 1) was detected on at least 1 day in 7
of the 11 mares (64%). Uterine edema was recorded on an average of 2.8 days (range 0±8
days) during the anovulatory wave. Before ®rst ovulation a large follicle (>25 mm) was
present for 12.4 days (range 7±19 days), and during this time edema was present for an
average of 3.1 days (range 0±8 days). Twelve of the 14 mares (86%) had at least 1 day of
uterine edema during follicular growth. By second ovulation a large follicle (>25 mm) was
present for a mean of 12.7 days (range 7±20 days). During this time edema was observed
for an average of 4.5 days (range 2±8 days). All 11 mares (100%) showed uterine edema at
some time during follicle growth.
Ten (91%) of the 11 mares that had serial large anovulatory follicles, had uterine edema
at some point prior to their ®rst ovulatory estrus of the breeding season. The edema was ®rst
detected 16±86 days before ®rst ovulation (mean ˆ 43  6:7 days before ®rst ovulation).
The ®ve mares that did not have anovulatory follicle waves did not have uterine edema
prior to ®rst ovulatory estrus. A large follicle tended to be present longer during an
anovulatory wave but this was not statistically different from the later waves because of the
high variability in persistence of the anovulatory follicles. There was no signi®cant
difference among follicle wave types in the length of time that uterine edema was present
(Table 1). However, uterine edema was present for a signi®cantly greater proportion of time
 E.D. Watson et al. / Theriogenology 59 (2003) 1181±1187 1185

Fig. 1. Anovulatory follicle (left; 45 mm in diameter) in the presence of estrous-like uterine edema (right) in a
mare, 37 days before the mare's ®rst ovulation.

(P < 0:05) in the presence of a large follicle at ®rst ovulation than during an anovulatory
wave, and was present relatively longest (P < 0:05) at second ovulation (Table 1).
During anovulatory follicle waves, estradiol concentrations were higher (P < 0:05) on
days when uterine edema was observed (6:5  1:6 pg/ml) than on consecutive days when
uterine edema was not present (5:5  1:5 pg/ml). Peak plasma estradiol concentrations and
mean estradiol concentrations during growth of the dominant follicle >25 mm were
signi®cantly higher (P < 0:001) during growth of follicles at ®rst and second ovulation
than during the growth of an anovulatory wave (Table 1). There was no signi®cant
difference in peak or mean estradiol concentrations between ®rst and second ovulations.
There were signi®cant correlations (P < 0:01) between mean and peak estradiol con-
centrations for each type of follicle wave.

4. Discussion

In the present study, we have shown that uterine edema occurs in the presence of
anovulatory follicles in pony mares, even though these follicles are producing very low
concentrations of estradiol. Most of the pony mares in the present study had uterine edema
at some point during growth of a large follicle, whether it was ovulatory or anovulatory, and
1186 E.D. Watson et al. / Theriogenology 59 (2003) 1181±1187

this was ®rst seen on average 43 days before ®rst ovulation. These ®ndings are in contrast
with those of Ginther [6] where estrous-like edema was not seen more than 15 days before
®rst ovulation. Another study in horse mares showed that estrous-like edema was present
periodically throughout spring transition, but this was not described in detail [1]. It has
been reported that the intensity of uterine edema is less at the ®rst ovulation of the breeding
season, compared with the second [6]. It is possible that there were increases in uterine
echogenicity within individual mares in the present study as the spring progressed.
However, on stud farms, where mares are often examined only periodically, subtle
differences within mares would not be detected. Therefore, it was considered to be valid
in the present study to record and report only the presence or absence of estrous-like edema.
It is known that large transitional follicles are steroidogenically incompetent [3,4]. In the
present study, plasma estradiol concentrations were low in the presence of anovulatory
follicles and only increased during growth of the ®rst ovulatory follicle. Based on
differences in uterine echogenicity, it has been suggested previously that estrogen
concentrations may be lower at the ®rst than at the second ovulation [6]. However, we
found no difference in peak or mean estradiol concentrations during growth of the
preovulatory follicle between ®rst and second ovulations, even though uterine edema
was present for a proportionately longer period of time (days of uterine edema/days follicle
present) during growth of the second preovulatory follicle, than the ®rst preovulatory
follicle.
Although estrous behavior was not monitored in the present study, it is well recognized
that transitional mares may remain in estrus for prolonged periods of time, despite low
circulating concentrations of estradiol. We observed small elevations in estradiol con-
centrations during the growth of anovulatory follicle waves, and circulating concentrations
were signi®cantly higher on the days that uterine edema was present than on the days
before and after detection of uterine edema. It seems likely, therefore, that in the absence of
progesterone, the mare is highly sensitive to very low concentrations of estrogen during
spring transition at the level of both the brain and the uterus. Indeed, a recent study has
suggested that progesterone is responsible for dissipation of uterine edema in the cyclic
mare [7]. Alternatively, another estrogen or metabolite could be in¯uencing uterine edema
or estrous behavior. In our study, there were highly signi®cant correlations between peak
estradiol concentrations and mean daily concentrations, which showed that either value can
be used to assess follicular steroidogenesis.
In our mares, anovulatory follicular waves tended to persist for a longer period of time
than ovulatory waves because they regressed rather than ovulating. Follicular regression
takes about twice as long as development [9]. Uterine edema was present for a similar
number of days in anovulatory waves and ®rst and second ovulatory waves. The presence
of estrous-like uterine edema during anovulatory waves means that edema cannot be used
as de®nite proof that a large follicle is destined to ovulate. However, because of the greater
persistence of anovulatory follicles, we showed that the chance of detecting uterine edema
in the presence of a large ovarian follicle was signi®cantly less in spring transition than
during cyclicity.
In conclusion, uterine edema was not a reliable indicator of follicular steroidogenic
competence during spring transition in pony mares, nor of whether the follicle would
subsequently ovulate.
 E.D. Watson et al. / Theriogenology 59 (2003) 1181±1187 1187

Acknowledgements

We thank Shirley Melling and Paul Wright for taking excellent care of the research
mares and Sheila Thomson for performing the estradiol assays. Reagents for the estradiol
assay were kindly provided by Professor S. Hillier, University of Edinburgh. This study
was supported by the Horserace Betting Levy Board.

References

[1] Ginther OJ. Folliculogenesis during the transitional period and early ovulatory season in mares. J Reprod
Fertil 1990;90:311±20.
[2] Davis SD, Grubaugh WR, Weithenauer J. Follicle integrity and serum estradiol 17b patterns during sexual
recrudescence in the mare. Biol Reprod Suppl 1987;36:121.
[3] Davis SD, Sharp DC. Intra-follicular and peripheral steroid characteristics during vernal transition in the
pony mare. J Reprod Fertil Suppl 1991;44:333±40.
[4] Watson ED, Heald M, Tsigos A, Leask R, Steele M, Groome NP, et al. Plasma FSH, inhibin A and inhibin
isoforms containing pro- and -aC during winter anoestrus, spring transition and the breeding season in
mares. Reproduction 2002;123:535±42.
[5] Roser JF, Valcheck DE, Lasley BL, Liu IK, Kojusner G, Taya K. Minimizing the incidence of endometritis
by predicting the ®rst ovulation of the year. Pferdeheilkunde 1997;5:539 [abstract].
[6] Ginther OJ. Uterus. Ultrasonic imaging and reproductive events in the mare. Cross Plains, WI:
Equiservices; 1986, p. 173±94.
[7] Pelehach LM, Sharp DC, Porter MB, Dixon LN, McDowell KJ. Role of estrogen and progesterone in the
control of uterine edema in pony mares. Biol Reprod Suppl 2000;62:386.
[8] Samper JC. Ultrasonographic appearance and the pattern of uterine edema to time ovulation in mares.
Proc AAEP, Phoenix 1997;43:189±91.
[9] Watson ED, Pedersen HG, Thomson SRM, Fraser HM. Control of follicular development and corpus
luteum function in the mare: effects of a potent GnRH antagonist. Theriogenology 2000;54:599±609.
[10] Palmer E. New results on follicular growth and ovulation in the mare. In: Roche JF, O'Callaghan D, editors.
Follicular growth and ovulation rate in farm animals. Dordercht: Martinus Nijhoff; 1987, p. 237±55.