hCG administration fell on cycle days 12ÔÇ“16. Six of the 17 cycles (35%)
156 PREVENTION OF OVARIAN HYPERSTIMULATION SYNDROME
produced viable pregnancies. All 17 patients developed signs of grade 2 or 3
OHSS but none developed severe OHSS. In 1995, the same authors treated
51 women at risk of developing OHSS by coasting and also waited until the
estradiol level dropped to below 3000 pg/ml (Sher et al., 1995). The clinical
pregnancy rate was 41% per oocyte retrieval (21 out of 51). None of the
patients developed severe OHSS; however, the mean number of embryos
transferred was 5.4 which was extremely high.
Ben-Nun et al. (1993) conducted a pilot study of 66 patients at risk of
developing OHSS. These patients were coasted and hCG was given when the
estradiol level reached 2500 pg/ml. Four of the 66 patients developed OHSS.
Finding a control group for patients undergoing coasting is not an easy
task. In an attempt to use an acceptable control group, Benadiva et al. (1997)
therefore compared coasting to cryopreservation. Gonadotrophins were
withheld in 22 patients at risk of OHSS. HCG was administered when the
estradiol levels dropped to 3000 pg/ml. The control group consisted of 26
patients in which no fresh embryo transfer was performed, and all the embryos
were cryopreserved and transferred during a subsequent unstimulated cycle.
Fertilization and delivery rates were not signi´¬ücantly different between the two
groups and Benadiva concluded that coasting could produce high pregnancy
rates without the need for multiple frozen/thawed cycles.
In a large retrospective study from Belgium, 120 women at risk of
developing OHSS were included (Dhont et al., 1998). These patients were
coasted when the estradiol levels exceeded 2500 pg/ml, and hCG was withheld
and then administered when the estradiol levels dropped below 2500 pg/ml.
The authors compared the outcome to those of 120 matched OHSS high-risk
patients who did not undergo coasting. Coasting signi´¬ücantly decreased the
incidence of moderate and severe OHSS.
Lee et al. (1998) carried out a pilot study of coasting in 20 patients at risk of
OHSS. The mean duration of coasting was three days. HCG was administered
on the day that the serum estradiol levels began to fall and four of the
20 patients developed severe OHSS despite coasting. The authors concluded
that the hCG administration was too early to prevent OHSS.
Tortoriello et al. (1998a) studied three groups of IVF patients. The ´¬ürst
group consisted of highly responsive coasted patients. The second group con-
sisted of equally responsive patients who did not undergo coasting. The control
group consisted of age-matched normally responsive patients. The rates of
moderate and severe OHSS did not differ statistically among the three groups.
No patient in group three developed OHSS. Moderate OHSS was diagnosed in
one patient from group one on the basis of sonographically demonstrable
minimal ascites. One patient in group two was a singleton pregnancy who
developed critical OHSS with severe hemoconcentration, oliguria, and a large
pleural effusion that required seven days of hospitalization.
Two subsets of coasted patients were also compared to assess the effect of
estradiol levels at the time they met the criteria for hCG. Subset one was
identical to group one, consisting of those 22 coasted patients who achieved
an estradiol level between 3000 and 3999 pg/ml at the time they met the
SECONDARY PREVENTION OF OHSS
criteria. Subset two consisted of the remaining 22 coasted patients excluded
from the comparison analysis who achieved estradiol levels of 44000 pg/ml
at the time they met the criteria. The two subsets did not differ regarding
age, FSH levels, instance of polycystic ovaries, number of oocytes retrieved or
oocyte maturity, fertilization and cleavage. These patients on average coasted
approximately one day longer than the less-responsive subset (p ┬╝ 0.0463).
The authors observed a signi´¬ücantly higher implantation rate, and the trends
suggested a higher clinical and multiple pregnancy rate. There were no signi´¬ü-
cant differences in severe or moderate OHSS among the two subsets. However,
all three of the patients who developed severe OHSS were in subset two, and
two of them were hospitalized for two days (Tortoriello et al., 1998a).
The estradiol and progesterone levels in all 44 coasted patients were
signi´¬ücantly reduced by the end of coast periods lasting longer than two days.
Linear regression analysis demonstrated a statistically signi´¬ücant positive
relationship between the duration of coasting and peak estradiol level achieved
(p < 0.0001), as well as a signi´¬ücant negative relationship between coasted
duration and the total number of mature oocytes retrieved (p ┬╝ 0.036). Logistic
regression analysis of coast interval duration also suggested an inverse relation
with the clinical pregnancy rate (p ┬╝ 0.09).
Tortortiello et al. (1998b) also observed severe OHSS despite coasting as
gonadotrophins were withheld when serum estradiol levels were 14 700 pmol/l.
An important but not surprising ´¬ünding was that there was a higher than
expected incidence of severe OHSS (33%) when coasting was started with
serum estradiol levels 429 000 pmol/l and a large number of follicles with
diameter larger than 18 mm.
In a study from Sweden, Waldenstrom et al. (1999) performed a multi-
center trial of coasting on 65 IVF cycles considered to be severely hyper-
stimulated. HCG was given when the estradiol levels fell below 10 000 pmol/l.
The mean duration of coasting was 4.3 days, 4 cycles were cancelled. The
pregnancy rate was 42% and the implantation rate was 31%, and only one
patient developed severe OHSS.
In an interesting Canadian study, Fluker et al. (1999) studied two groups
of IVF patients undergoing coasting with mature and immature follicles.
In the ´¬ürst group (n ┬╝ 63), estradiol concentration rose rapidly and exceeded
3000 pg/ml. Exogenous gonadotropins were withheld to allow estradiol
concentrations to decrease by at least 25% before hCG administration.
Each subject met the follicular criteria for hCG administration and oocyte
retrieval, ┬Ş3 follicles of ┬Ş18 mm before or during the coasting period. Oocyte
retrieval occurred 34 h later. The luteal phase was supported with micronized
progesterone at a dosage of 200 mg twice daily. In the second group
(n ┬╝ 30) estradiol concentrations rose rapidly in the presence of numerous
intermediate size follicles. In anticipation of overstimulation, the HMG dosage
was reduced. This was followed by an abrupt and inadvertent decline in
estradiol concentrations before the attainment of appropriate follicular
maturity. Gonadotrophin treatment was then reinstituted to restimulate
follicular growth. The hCG, 10 000 IU, was administered once three or more
158 PREVENTION OF OVARIAN HYPERSTIMULATION SYNDROME
follicles ┬Ş3 mm were achieved. Oocyte retrieval was performed as per the
routine protocol. The mean age, etiology and duration of infertility was similar
between the two groups. The average duration of rise was longer in the
´¬ürst group, 3.3 days, than in the second group, 1.37 days, in keeping with the
larger follicles and more established steroid oogenesis (Fluker et al., 1999).
Clinical pregnancies occurred in 23 of the 63 cycles (36.5%) and in 12 of
30 cycles (40%). The implantation rate per embryo was 14.3% and 17.8%,
respectively. Eleven women (12%) had evidence of moderate OHSS, which was
managed conservatively at home. One woman (1.1%) from the second group
was hospitalized with severe OHSS that required treatment with paracentesis
and intravenous albumin. The authors noted that their implantation rates
compared favorably with those of Sher et al. (1993), 15.4% versus 9.5%,
respectively, as did the clinical pregnancy rates, 37.5% versus 41%, respectively,
despite the difference in the number of embryos transferred, 2.9 versus 5.4,
respectively. The authors suggested that IVF cycles do not have to be markedly
overstimulated to have enough reserve to withstand the coasting period.
Rather, a limited period of coasting before the administration of hCG may
improve the margin of safety and still be well tolerated, even in cycles in which
the response is only slightly increased. The approach in the IVF cycles in
the study by Fluker et al. (1999) was different from those in the studies by
Sher et al. (1993, 1995). In the two clinical studies by Sher et al. (1993, 1995)
the estradiol levels were 46000 pg/ml, or had 430 follicles and received
hCG after their estradiol levels decreased to <3000 pg/ml. In contrast, the
more conservative approach to stimulation in the study by Fluker et al.
(1999), resulted in lower peak estradiol levels and a less precipitous decline
in estradiol concentration. As a result, only 18 of 93 patients undergoing IVF
in their study had estradiol levels of 46000 pg/ml and 28 received hCG, even
though their estradiol levels remained 43000 pg/ml. Fluker et al. (1999) high-
lighted that the only women in whom severe OHSS developed in their study
had an estradiol level of 2762 pg/ml, which is below the limit suggested by most
investigators as a safe point to give hCG and have success with coasting.
Aboulghar et al. (2000) performed a ÔÇ˜ÔÇ˜hybridÔÇ™ÔÇ™ study. In a prospective
randomized study the authors evaluated the incidence of OHSS in 49 high-risk
patients using a reduced hMG dose in one arm and continuation of the same
dose in the other arm before coasting. There were no cases of severe OHSS in
either group after coasting; however, the duration of coasting was signi´¬ücantly
reduced when the dose of hMG was reduced. A historical control group was
used to compare the two subsets of the coasted patients. The incidence of severe
OHSS in the historical control group was 25%, as compared with 17% in the
coasted group (Aboulghar et al., 2000).
In a study from Japan, Ohata et al. (2000) performed coasting in ´¬üve
patients with PCOS who had been previously hospitalized due to severe OHSS
in a previous IVF cycle. Coasting was effective in preventing OHSS in these
Grochowski et al. (2001) performed a coasting study on 112 hyperstimu-
lated IVF patients when the estradiol level was over 3000 pg/ml and the leading
SECONDARY PREVENTION OF OHSS
follicleÔÇ™s diameter was ┬Ş18 mm. Fertilization failed in six patients. All the
embryos were frozen in another 10 patients. The pregnancy rate was 30.4% and
the implantation rate was 18.1%. Moderate OHSS occurred in six patients
and severe OHSS occurred in another two patients.
Al-Shawaf et al. (2001) performed a modi´¬üed coasting protocol in patients
at risk of severe OHSS based on ultrasound monitoring. Serum estradiol
levels were measured only in patients with 420 follicles on ultrasound.
Moderate OHSS occurred in three patients (0.7%) and severe OHSS in one
patient (0.2%). Pregnancy rates were 39.6% and 40% in cycles where the
gonadotrophin dose was reduced or withheld, respectively.
Al-Shawaf et al. (2002) determined that measuring serum FSH in addition
to estradiol levels during coasting could assist in predicting the point at which
the serum estradiol level had declined to a suf´¬üciently safe point for hCG
When to Initiate Coasting?
Egbase et al. (2002) performed a pilot study to determine the impact of
withholding gonadotrophins at an earlier stage in patients at risk of developing
OHSS. The background of the study was their observation that OHSS still
occurred despite coasting in patients with excessive follicular response and high
estradiol levels. In this trial, they withheld gonadotrophins for a ´¬üxed period
of three days once the lead follicle was 15 mm with continuation of pituitary
downregulation. They investigated 102 obese patients with PCOS in whom
there was evidence of excessive follicular response, more than 10 follicles per
ovary and serum estradiol levels 41500 but <3000 pg/ml. The mean serum
estradiol level on coasting day 1 was 1943 pg/ml and 2169 pg/ml on the day of
hCG administration. Normal fertilization and cleavage rates were observed
and the clinical pregnancy rate was 45%. There were no cases of severe OHSS.
Four patients suffered pregnancy-associated late-onset moderate OHSS. The
authors concluded that early withholding of gonadotrophins in patients with
excessive follicular response at risk of developing severe OHSS is consistent
with good embryological and clinical outcome in IVF cycles.
VERY EARLY COASTING
In a recent study presented at the American Society for Reproductive Medicine,
very early coasting was attempted in order to prevent OHSS (Lukaszuk et al.,
2005). The authors studied 27 patients who had experienced OHSS in previous
ICSI cycles. For the next ICSI procedure, they were randomized into two groups:
12 patients underwent ovarian stimulation and received 225 IU hMG for
two days followed by two days without hMG and then resumed stimulation.
The second group received standard doses of stimulation without interruption
of their hMG injections. The authors observed that very early coasting resulted
in no cases of OHSS compared to six cases in the control group. They suggested
that a multi-center trial could determine the value of this approach.
160 PREVENTION OF OVARIAN HYPERSTIMULATION SYNDROME
Levinsohn-Tavor et al. (2003) appraised the three factors that should be
considered for the initiation of coasting. Plasma estradiol concentration, which
re´¬‚ects the total functional granulosa cell population, the number of ovarian
follicles, which predicts the potential for further granulosa cell population and
estradiol rise, and, ´¬ünally, the diameter of the leading follicles, are the three key
Most publications addressing coasting show that an estradiol concen-
tration of 2500 to 3000 pg/ml was the value most commonly chosen by the
clinicians (Sher et al., 1995; Benadiva et al., 1997; Dhont et al., 1998; Lee et al.,
1998; Tortoriello et al., 1998; Fluker et al., 1999; Al-Shawaf et al., 2001).
The relatively low threshold for coasting has been shown to effectively
reduce the incidence of OHSS without compromising the cycle outcome
(Levinsohn-Tavor et al., 2003) High cutoff levels around 6000 pg/ml are
associated with a higher incidence of OHSS and the need for longer periods of
coasting (Waldenstrom et al., 1999).
Since the ´¬ürst study by Rabinovici et al. (1987), it has been demonstrated
that, even after withholding gonadotrophins, a subsequent rise in serum
estradiol for one or two days still occurs (Sher et al., 1995; Fluker et al., 1999;
Egbase et al., 2000) (Figure VII.10). In one of the earlier studies when coasting
was initiated at a plasma estradiol value of over 3000 pg/ml, the plasma
estradiol increased to over 6000 pg/ml during the coasting period.
WHEN TO ADMINISTER hCG AND END COASTING?
The timing of hCG administration at the end of coasting is a matter of
great clinical signi´¬ücance in terms of avoidance of OHSS and successful
outcome. Administration of hCG when the estradiol level drops below 2500
to 3000 pg/ml has been termed to be effective in lowering the risk of OHSS
(Sher et al., 1995; Benadiva et al., 1997; Dhont et al., 1998; Tortoriello et al.,
1998; Al-Shawaf et al., 2001). Dhont et al. (1998) compared a coasting group
with a control group. Both had a similar maximum estradiol level (3830 pg/ml)
and number of follicles (n ┬╝ 24). On the day of hCG administration estradiol
levels were 2348 pg/ml in the coasted group compared with 3833 pg/ml in the
control group. Only one patient in the coasted group developed severe OHSS
compared with nine patients in the control group.
Levinsohn-Tavor et al. (2003) advised that when an appropriate threshold
for administering hCG is attained, serum estradiol should be followed and not
allowed to fall too low below the threshold. I fully agree with this excellent
clinical remark. However, the greatest dif´¬üculty that we have encountered is the
signi´¬ücant drop in the estradiol level that can occur in one day. Our group
encountered two patients whose estradiol level was above 4000 pg/ml during
coasting and dropped to below 1000 pg/ml over a 24 h period, with a
detrimental effect on the quality of oocytes and pregnancy outcome (Grace
et al., 2005). Waldenstrom et al. (1999) reported two cases in which they
delayed hCG for an additional two days after the serum estradiol level had
dropped below the threshold level of 2724 pg/ml, which led to bleeding and
cycle cancellation. In three other cases, serum estradiol was allowed to fall
SECONDARY PREVENTION OF OHSS
below the threshold value, resulting in the retrieval of one to three oocytes
of poor quality.
THE DURATION OF COASTING
How many days of coasting can be carried out without compromising the
outcome of ovarian stimulation? The number of recorded days of coasting
has varied between 1 and 11 days (Table VII.10). The effect of the duration
of coasting has remained controversial. While some studies suggested that
gonadotrophins could be withheld for 10 or more days without compromise
of the outcome, others have reported a decrease in pregnancy rate when the
duration has exceeded four days (Tortoriello et al., 1998a; Ulug et al., 2002;
Isaza et al., 2002; Grace et al., 2005).
Ulug et al. (2002) carried out a retrospective study to de´¬üne the optimal
interval of coasting in patients at high risk of developing OHSS. In their study,
patients were characterized according to the number of days between the cessa-
tion of gonadotrophins and hCG administration. Overall, out of 207 patients
coasted, coasting lasted one day in 39 cycles, two days in 61 cycles, three days in
49 cycles, and ┬Şfour days in the remaining 59 cycles. Patients in whom coasting
lasted ┬Şfour days had signi´¬ücantly reduced implantation and pregnancy rates
compared to patients with a shorter coasting interval. The authors concluded
that coasting for more than four days appears to reduce implantation and
pregnancy rates where oocyte and embryo quality did not appear to be affected.
The authors suggested that, in patients who needed coasting for more than three
days, cryopreservation of the embryos should be considered.
Moreno et al. (2004) performed a retrospective study of 132 patients
who demonstrated a high response to ovarian stimulation with estradiol
44500 pg/ml and/or more than 20 follicles 417 mm, and who were coasted
due to the high risk of developing OHSS. The authors investigated the impact
of the duration of coasting on IVF cycle outcome. In addition, serum
progesterone and LH were measured to investigate whether premature
luteinization was present in the cycles, and whether it might be related to
coasting duration. A signi´¬ücant difference in implantation rate was observed
when coasting was required for more than four days, together with a trend
towards a high cancellation rate. Premature luteinization was signi´¬ücantly
elevated in women undergoing coasting compared with control women,
34 versus 15.6% (p < 0.05). In the majority of patients who showed premature
luteinization, coasting lasted three days. The authors concluded that prolonged
coasting may affect the endometrium in relation to the implantation window.
These ´¬ündings may explain why some patients undergoing extended coasting
demonstrate a lower implantation rate compared with controls.
COASTING IN GnRH ANTAGONIST IVF CYCLES
The use of GnRH antagonists has simpli´¬üed the IVF cycle for many patients
who were not used to using GnRH agonist from the luteal phase of the cycle
preceding their IVF cycle. Most of the data on coasting in IVF come from the
GnRH agonist long protocol, but as the use of the GnRH antagonist becomes
162 PREVENTION OF OVARIAN HYPERSTIMULATION SYNDROME
more commonplace, cases of coasting in GnRH antagonist cycles are being
reported. Delvigne et al. (2001) reported two cases in which coasting was used
in a stimulation regimen that included GnRH antagonist and gonadotrophins.
The ´¬ürst case had an estradiol level of 7851 pg/ml on day 16 of the cycle, and the
second patient had a level of 6701 pg/ml on day 13 of her cycle. The ´¬ürst patient
increased her level of estradiol on the ´¬ürst day after coasting and had a rapid
and clinically signi´¬ücant 83% decrease in the level of estradiol. The second
patient experienced a more progressive decrease of estradiol, but her estradiol
level did not increase after hMG administration was stopped. Neither patient
developed OHSS. The authors suggested that coasting could be used when the
stimulation involves hMG and GnRH antagonist.
Another interesting case comes from Belgium where a 23-year-old patient
with PCOS was referred to the Center for Reproductive Medicine in Brussels
because of a high risk of developing OHSS. She exhibited a rising LH following
ovulation induction with a low-dose step-up protocol using gonadotrophins
(Fatemi et al., 2002). The patient was counseled to perform a rescue IVF cycle
and was coasted using 0.25 mg of ganirelix. The serum estradiol concentrations
decreased and the LH peak was successfully suppressed. No OHSS occurred and
a twin pregnancy resulted after the transfer of two embryos.
HOW SUCCESSFUL IS COASTING IN ELIMINATING OHSS?
Coasting is very successful in decreasing if not eliminating OHSS (Grace et al.,
2005). There is no question that coasting has had a tremendous impact on
the clinical management of OHSS. Most investigators have included it at least
as part of their approach to prevent this syndrome. From an evidence-based
medicine point of view, there is a sparcity of data in terms of prospective
randomized trials. The reason for this is obvious, I and others believe it would
be dif´¬ücult to have a randomized clinical trial in which one of the arms would
not be coasted and therefore subjected to the risk of severe OHSS. Therefore,
many investigators have included a control group or perhaps another modality
of OHSS prevention (Rizk, in press).
DÔÇ™Angelo and Amso (2002a) performed a Cochrane review on coasting for
the prevention of OHSS. They identi´¬üed 13 studies but only one trial met
the inclusion criteria, therefore it was concluded that there was insuf´¬ücient
evidence available to determine whether coasting was an effective strategy in
preventing OHSS. In the only prospective study, 15 patients were included in
each study arm comparing coasting with unilateral follicular aspiration,
a technique that is seldom used. The Cochrane review stressed the absence
of high-quality studies, which limited to a great extent the conclusions that could
IS THERE A PROBLEM WITH COASTING?
The greatest concern about adopting a policy of coasting or withholding
gonadotrophins is a decrease in the quality or number of oocytes and a
SECONDARY PREVENTION OF OHSS
subsequent drop in the pregnancy rate. Aboulghar et al. (1997) studied oocyte
quality in patients with severe OHSS. They reported that the inferior quality
and maturity of oocytes in OHSS reduced the fertilization rate, but did not
affect the quality or the number of embryos transferred, or the pregnancy rate.
The effect on oocyte quality could be due to the prevalence of polycystic ovaries
in this group of patients.
The literature is divided between those who believe that coasting has some
negative impact on oocyte quality (Grace et al., 2005; Rizk, in press) and those
who believe that coasting has no impact on the quality of oocytes (Delvigne
et al., 2002; Isaza et al., 2002). The criteria used for the initiation and
determination of coasting that has been published in the literature is very
heterogeneous. This probably explains some of the differences between the
reports. In 2002, Ulug et al. found that coasting for four or more days reduces
implantation and pregnancy rates, while oocyte quality does not appear to be
affected. Isaza et al., in 2002, compared cycle outcome in recipients of oocyte
donation from donors who underwent coasting and donors who did not. The
outcome of oocyte donation from donors undergoing coasting was not
impaired, with similar implantation and pregnancy rates. If the duration of
coasting was longer than four days, a signi´¬ücant decrease in implantation and
pregnancy rates was found. Delvigne et al. (2002) performed a retrospective
cohort study of 157 patients compared with a controlled group of 208 IVF
cycles, which had reached serum estradiol levels of at least 4000 pg/ml without
being coasted. In the group of coasted cycles, the question of whether indirect
parameters related to coasting had an effect on IVF outcome was also analyzed.
The authors observed that patients who had undergone coasting had higher
maximum estradiol levels and greater numbers of large follicles (p < 0.001)
and lower oocyte recovery rates (p < 0.001) than the control group. The IVF
outcomes were similar between the two groups. The authors also observed that,
within the group of patients who had undergone coasting, no signi´¬ücant
relationship was observed between the number of coasting days, the estradiol
levels on the day of hCG, or the fallen estradiol level and the outcome, whether
measured in terms of oocyte quality, pregnancy rate or OHSS occurrence.
In a recent study we have observed a decline in the implantation and