Grade
Symptoms
Mild OHSS
Abdominal bloating
Mild abdominal pain
Ovarian size usually <8 cma
Moderate OHSS
Moderate abdominal pain
Nausea ± vomiting
Ultrasound evidence of ascites
Ovarian size usually 8–12 cma
Severe OHSS
Clinical ascites (occasionally hydrothorax)
Oliguria
Haemoconcentration haematocrit >45 %
Hypoproteinaemia
Ovarian size usually >12 cma
Critical OHSS
Tense ascites or large hydrothorax
Haematocrit >55 %
White cell count >25,000/ml
Oligo/anuria
Thromboembolism
Acute respiratory distress syndrome
Ovarian diameter has historically been used as an indicator of the severity of OHSS, but the evidence linking ovarian size with pathophysiological derangements is poor. If ovarian size is considered, a modern approach could be based on ovarian volume, rather than diameter. However, it is not known whether ovarian volume is related to the degree of fluid shift, and whether the total volume of both ovaries should be considered rather than individual ovarian volume. At present we simply do not know enough about this subject to be able to use ovarian volume to classify the severity of OHSS.
In addition, the time of onset of OHSS should be considered when classifying this condition. “Early” OHSS occurs within 9 days of hCG administered for final follicular maturation and is precipitated by the effect of exogenous hCG. “Late” OHSS occurs 10 or more days after the ovulatory dose of hCG and, is precipitated by endogenous hCG from an early pregnancy. “Late” OHSS is significantly more likely to be severe than is “early” OHSS [11].
Prediction of OHSS
The incidence of moderate or severe OHSS has been reported to lie between 3.1 and 8 % of IVF cycles [14]. Much less commonly, OHSS may develop following ovulation induction with gonadotrophins or clomifene. A study utilizing hospital discharge statistics from Finland in the period between 1996 and 1998 identified hospitalization due to OHSS following 0.04 % of ovulation induction cycles and 0.9 % of IVF cycles [15].
Pre-treatment Patient Characteristics
Young age, a previous history of OHSS, and the presence of polycystic ovaries (PCO) or polycystic ovarian syndrome (PCOS) have been shown to increase the risk of developing OHSS [14]. VEGF activity is increased in the theca cells of women with PCOS and they may also have a larger number of theca cells compared to women with morphologically normal ovaries.
Serum anti-mullerian hormone (AMH) concentration is an accurate measure of ovarian reserve and has been studied as a predictor of the risk of OHSS. A number of studies have found that high serum AMH levels are predictive of the risk of over-response and OHSS [16, 17]. As a result of these studies, many clinicians now favor AMH-based tailored ovarian stimulation protocols, aiming to optimize ovarian stimulation and avoid the extremes of ovarian response. One of the benefits of such an approach is a possible reduction in the incidence of OHSS, although randomized controlled data for this effect are lacking. Yates et al. [18] and Nelson et al. [19] described their results with using serum AMH to select women with a high risk of developing OHSS for GnRH antagonist protocols with a starting dose of 150 iu FSH daily. Nelson et al. carried out a prospective cohort study between two centers, one of which used GnRH agonist and the other GnRH antagonist to control LH during ovarian stimulation with high AMH concentrations. Hospitalization for OHSS was required in 20 out of 148 women receiving GnRH agonist (13.9 %) compared to 0 out of 34 women receiving GnRH antagonist, albeit in a different center. The authors concluded that pretreatment serum AMH levels could be used to individualize the treatment regime and thereby reduce the risk of OHSS. Yates et al. compared retrospectively 346 cycles of IVF based on a conventional protocol using serum FSH and age to guide stimulation, with 423 cycles using an AMH-guided stimulation protocol. In the conventional group, long protocol GnRH agonist was used in most cases, while, in the AMH-guided group, women with high AMH levels received GnRH antagonist and 150 iu FSH daily. The introduction of AMH-guided ovarian stimulation was associated with a reduced incidence of cycle cancellation or ‘freeze-all’ due to a perceived risk of OHSS, although the incidence of hospital admission due to severe OHSS did not differ significantly between the two groups [19].
The antral follicle count (AFC) on transvaginal ultrasound scan is another measure of ovarian reserve which correlates with the ovarian response to stimulation. A high AFC is associated with an increased risk of developing OHSS. Jayaprakasan et al. found that the incidence of OHSS was 2.2 % in women with an AFC <24 and 8.6 % in women with AFC ≥24 [20]. The efficacy of AFC in predicting excessive ovarian response (defined as the collection of 20 or more oocytes) was identified by Kwee et al. [21] who found that an AFC of 14 yielded the best combination of sensitivity (82 %) and sensitivity (89 %) with a positive predictive value of 58. AFC has been proposed as a guide to individualizing ovarian stimulation regimes [22] by identifying patients who may over-respond to stimulation, allowing the choice of an appropriate regime.
It is clear that AMH and AFC are useful markers of how the ovaries will respond to stimulation, but it should be kept in mind that around one-third of cases of severe OHSS occur in cycles where the ovarian response was not considered excessive. This is an inherent limitation of using a marker of ovarian response to eliminate the risk of OHSS.
Ovarian Response Parameters
Once ovarian stimulation has commenced, monitoring of specific parameters of response may help to identify cycles at an increased risk of developing OHSS. These include high serum concentrations of oestradiol (E2) large numbers of follicles, and a large number of oocytes retrieved [14]. The desire to identify cycles with over-response and thereby identify those that might result in OHSS is one of the main reasons why stimulated IVF cycles are monitored closely in clinical practice. However, evidence shows that ovarian response parameters have only modest predictive value for OHSS. A significant proportion of cases of severe OHSS occur in cycles where no risk factors were identified in the antecedent treatment cycle or patient characteristics, while the incidence of severe OHSS in cycles considered “high-risk” by commonly used predictive variables is around 20 %.
Further, there is no clear cut-off level of different ovarian response parameters that allows us to reliably differentiate “high” from “low” risk cycles. This is probably because the risk of developing OHSS lies along a continuum. For instance, there is a certain risk of developing OHSS in a cycle where 20 eggs are collected, but even with the collection of fewer eggs there is a risk of OHSS, albeit lower than with 20 eggs. Hence, any treatment cycle in which supraphysiological ovarian stimulation is used should be considered at risk of OHSS. Where ovarian response parameters are used to target preventative measures, it should be recognized that these parameters are unable to predict all cases of severe OHSS.
Preventative Measures
Alternatives to Gonadotrophins
The risk of OHSS inherent in the use of gonadotrophins means that clinicians should always consider alternatives where appropriate. Women with ovulatory dysfunction should be advised about lifestyle modification to optimize weight or avoid excessive exercise as the first step. In women with PCOS, clomifene, metformin, aromatase inhibitors, and laparoscopic ovarian diathermy should be considered before IVF. This may require careful patient counselling and clinicians should resist the temptation to see IVF as an easy solution where other, safer options are applicable.
In vitro maturation (IVM) of oocytes obtained from unstimulated ovarian follicles is an alternative to ovarian stimulation with gonadotropins for some women with PCIS. IVM has a lower likelihood of leading to live birth than stimulated IVF [23], but it is likely that success rates will improve with advances in technique and knowledge.
Prevention of OHSS in IVF Treatment
Laparoscopic Ovarian Diathermy
A controlled trial in women with ultrasound evidence of PCO found a lower risk of cancellation due to over-response, but no difference in the incidence of OHSS, with the use of laparoscopic ovarian diathermy carried out 1 week prior to the start of gonadotrophin stimulation in GnRH agonist cycles [24]. At present, there is no clear indication to perform the procedure prior to IVF with the aim of reducing the risk of OHSS.
Ovarian Stimulation Regimes—Starting Dose of FSH
It is logical to base the starting dose of FSH for controlled ovarian hyperstimulation on the ovarian reserve and the likelihood of an excessive ovarian response. AMH and AFC are now widely used for this purpose, along with other factors such as the presence of PCO and a previous history of OHSS. There are no randomized trials examining a lower starting dose of FSH for patients with risk factors for excessive ovarian response undergoing IVF. Marci et al. observed a significant reduction in cycle cancellation rate with a starting dose of 75 iu rFSH in 61 women who had previously responded excessively to a starting dose of 150–225 iu hMG. The incidence of OHSS was not reported in the study, but commonly used risk parameters (peak oestradiol and number of oocytes) were reduced in the group with the lower starting dose [25].
Choice of FSH
The risk of OHSS is not altered by the type of FSH used (urinary vs recombinant) [26].
GnRH Agonist Versus GnRH Antagonist
The incidence of OHSS is significantly lower in cycles where a GnRH antagonist is used, compared to cycles using GnRH agonist to effect pituitary down-regulation. Al-Inany et al. performed a meta-analysis of randomized trials comparing GnRH antagonist with long protocol GnRH agonist treatment cycles [27]. Information on the incidence of severe OHSS was available in 29 studies with a total of 5,417 subjects. The incidence of severe OHSS was significantly lower in GnRH antagonist cycles compared to GnRH agonist cycles (2.65 % vs 6.61 %; 95 % CI = −0.05 to −0.02; p < 0.00001). In the overall study population, the risk of OHSS was 60 % lower in women receiving GnRH antagonist, with an absolute risk reduction of 4 % and a corresponding number needed to harm of 25 (that is, for every 25 women receiving the long protocol GnRH agonist regime, there would be one extra case of severe OHSS). The protective effect of GnRH antagonist was even more marked when women with PCOS were considered separately: the incidence of severe OHSS among women with PCOS was significantly lower with the use of GnRH antagonist (3.44 % vs 15.02 %; 95 % CI = −0.14 to −0.07; p < 0.00001). The incidence of coasting or cycle cancellation due to a perceived risk of OHSS was also significantly lower in GnRH antagonist cycles, lending further plausibility to the findings.
Despite this evidence of the benefit of GnRH antagonist in reducing the risk of OHSS, some clinicians remain wary of using it widely. This is partly due to a fear that IVF live birth rates may be lower if they change to a new regime, owing to the existence of a ‘learning curve’ in the application of GnRH antagonist [28, 29].
Modifying the “Trigger”—GnRH Agonist for Final Follicular Maturation in GnRH Antagonist Cycles
The risk of OHSS in GnRH antagonist cycles may be further reduced by exploiting a specific feature of the antagonist: in women administered GnRH antagonist, pituitary gonadotroph cells remain sensitive to GnRH. As a result, GnRH agonist administration to a woman receiving GnRH antagonist leads to an initial ‘flare’ effect, characterized by the release of LH and FSH. The resulting LH surge is sufficient to cause follicular maturation, without the need for an exogenous hCG trigger. LH has a shorter half-life than hCG and is less likely to precipitate OHSS than hCG [30].
GnRH agonist trigger for final follicular maturation has been studied in several trials and appears to significantly reduce the risk of developing OHSS compared to an hCG trigger in GnRH antagonist cycles. A meta-analysis of 5 randomized controlled trials comprising 504 fresh autologous cycles found a significantly lower risk of OHSS with a GnRH agonist trigger compared with HCG trigger (OR 0.10, 95 % CI 0.01–0.82), suggesting that for a population with an OHSS incidence of 3 % using HCG trigger, the incidence using GnRH agonist trigger would be 0–2.6 %. The incidence of OHSS in egg donation cycles (3 trials, 342 cycles) was also significantly lower with GnRH agonist trigger compared to HCG trigger (OR 0.06, 95 % CI 0.01–0.31) [31].
Disappointingly, the reduction in OHSS risk with this approach carries a significant downside: safety is purchased at the expense of a lower likelihood of success. Meta-analysis shows significantly lower live birth rates (OR 0.44, 95 % CI 0.29–0.68; 4 trials comprising 497 cycles) and increased risk of miscarriage (OR 1.89, 95 % CI 1.11–3.21; 8 trials comprising 713 cycles) with GnRH agonist trigger compared to HCG trigger in autologous IVF cycles. The most likely reason for this is that the endogenous LH surge induced by the agonist trigger is of a shorter duration than a typical spontaneous pre-ovulatory LH surge. While this attenuated surge is sufficient for oocyte maturation, it is not adequate for normal corpus luteum formation, leading to a reduced implantation rate and a higher rate of early pregnancy loss.
This drawback limits the use of agonist trigger in clinical practice and attempts have been made to try and overcome the problem of luteal defect associated with this treatment. One approach, somewhat paradoxically, is to administer a small dose of HCG either at the time of GnRHa trigger (so called ‘dual trigger’) or at the time of egg retrieval. In a controlled trial of 302 patients, Humaidan et al. found that 1,500 iu HCG administered 35 h after the GnRHa trigger was associated with no cases of OHSS and live birth rates comparable to a conventional HCG trigger in unselected women undergoing IVF with a GnRH antagonist regime [32]. However, a retrospective study of 23 women at increased risk of OHSS found severe early OHSS in 6 patients (26 %) with the use of GnRH agonist trigger and 1,500 iu HCG luteal rescue [33]. HCG administration, even in small doses, in women at high risk of OHSS remains problematic. Alternatives to this include intensive steroid replacement (intramuscular progesterone and transdermal or oral oestradiol) [34, 35], recombinant LH injections during the luteal phase [36], and daily intranasal GnRH agonist during the luteal phase [37]. At present there is insufficient evidence to say which, if any, regime is able to overcome the luteal phase defect, and further research is clearly needed. However, it is clear that if fresh embryo transfer is not planned, and in oocyte donation cycles, GnRH agonist trigger is preferable to hCG trigger because of superior safety.
Recombinant LH/hCG
The observation that hCG has a stronger stimulatory effect on granulosa cells led to an expectation that recombinant LH for follicular maturation would be associated with a lower risk of OHSS than hCG. However, this was not borne out by a meta-analysis of two randomized trials. Similarly, recombinant hCG carries the same risk of OHSS as urinary hCG when used as the trigger.
Dose of hCG
Given the correlation between degree of HCG exposure and the risk of OHSS it appears sensible to use the lowest effective dose of HCG for final follicular maturation in cycles considered at risk of OHSS. For urinary hCG a dose of 2,500 iu may be sufficient in women with a high ovarian response [38]. A dose of 250 mcg recombinant hCG is associated with a lower risk of OHSS than a dose of 500 mcg [39].
Coasting
‘Coasting’ refers to the practice of withholding gonadotrophins while maintaining pituitary suppression. Daily serum E2 estimation and follicular tracking are carried out until E2 drops to a “safe” level. HCG is then administered, followed by oocyte retrieval and embryo transfer. Serum FSH levels decline significantly during the coasting period. Larger follicles have a lower dependence on FSH than smaller follicles and are capable of continuing their growth and maturation, while small and intermediate follicles undergo atresia [40, 41]. Granulosa cell apoptosis may increase as FSH concentrations fall, accompanied by falling concentrations of vaso-active mediators produced by the ovaries [42, 43].
Although coasting is widely used, it has not been subjected to adequate prospective research comparing it to a control strategy. A Cochrane review [44] identified four randomized trials of coasting, but only one of these compared coasting with no coasting, showing a reduction in the risk of moderate or severe OHSS with coasting (OR 0.17, 95 % CI 0.03–0.88; P = 0.03). Other trials either compared coasting with early follicular aspiration or GnRH antagonist administration, and no significant difference was found in the incidence of OHSS in these comparisons. Owing to the comparison with another method of prevention, these trials may not be a true reflection of the value of coasting. There are several retrospective studies examining the value of coasting in preventing OHSS. It is clear that coasting does not abolish the risk of OHSS, but there does appear to be a lower incidence of OHSS in coasted cycles than would be expected from the literature [45].
Coasting can be initiated once follicles are mature (>15 mm). The criteria described for initiating coasting are variable, but are intended to identify cycles where the ovarian response is so excessive that cycle cancellation would otherwise be considered. Most investigators have recommended using E2 levels as one of the criteria for initiating coasting, with cut-off levels varying from 2,500 pg/ml [41] to 6,000 pg/ml [46]. The experience of the individual center and an overall assessment of the risk of OHSS in any given treatment cycle should be taken into account. The number of follicles is also a criterion described by most investigators, with some variation in the cut-off numbers used and whether the total number of follicles or the number of mature follicles only is considered.
Criteria for when coasting may cease, followed by the administration of hCG, usually relate to a drop in serum E2 level to a “safe” level, usually below 3,500 pg/ml [45] or 3,000 pg/ml. Mansour et al. reported their experience of a large retrospective series of cycles in which coasting was initiated if there were ≥20 follicles and E2 ≥ ,000 pg/ml. Of 1,223 cycles with coasting, 16 cases of severe OHSS occurred (1.3 %), all in cycles where hCG was administered prior to the E2 dropping below 3,000 pg/ml [47].
Avoiding hCG
Cycle Cancellation
OHSS develops only in cycles with exposure to hCG or LH. As a result, if hCG is withheld in cycles at risk of OHSS and an endogenous LH surge is avoided, OHSS should not develop. The emotional and financial costs of cancelling a treatment cycle in this way can be significant, making patients and clinicians understandably reluctant to let the ovarian response go to waste. However, there are often situations where cycle cancellation is the safest approach and patients can be counseled about starting afresh with a modified treatment regime, using GnRH antagonist and a lower starting dose of FSH.
Cryopreservation of All Embryos
Avoiding fresh embryo transfer eliminates exposure to endogenous hCG of pregnancy and should thereby eliminate the possibility of late OHSS. Clearly, it cannot prevent early OHSS which is related to the pre-ovulatory exogenous HCG. Endo et al., in a controlled trial of 138 women undergoing elective cryopreservation of all embryos due to excessive ovarian response, found that the incidence of OHSS was reduced by the continuation of GnRH agonist (avoiding an endogenous LH surge) for 1 week after hCG injection [48].
Adjuvant Treatments
Metformin Co-treatment During Gonadotrophin Stimulation
Metformin has been shown to be useful in preventing OHSS in women with PCOS undergoing IVF. A typical regime would be to use Metformin 850 mg twice daily from the first day of down-regulation to the day of oocyte retrieval [49]. A systematic review pooling data from five randomized controlled trials in women with PCOS undergoing IVF showed a significant reduction in the incidence of OHSS with the use of metformin (13/227 vs 47/222; OR 0.27, 95 % CI = 0.16–0.47) [50]. There was no statistically significant difference in markers of ovarian response or treatment outcome with the use of metformin. Only one trial (Doldi et al. 2006) examined the use of metformin in combination with a GnRH antagonist regime, studying 40 women with PCOS randomized to no pre-treatment or metformin 1.5 g daily for 2 months prior to the start of IVF. The incidence of OHSS was 5 % in the metformin group vs 15 % in the placebo group (p < 0.05) [51].
Dopamine Agonists
Dopamine agonists have been proposed as a preventative measure for OHSS, based on the action of dopamine in antagonising the vascular permeability-enhancing effect of VEGF through the dopamine receptor type 2. Initial studies in rats were followed by a trial in oocyte donors which showed a reduced incidence of moderate, but not severe, OHSS in oocyte donors receiving 0.5 mg Cabergoline daily from the day of HCG administration for 8 days [52]. It appears that cabergoline is effective in preventing early, but not late OHSS [53], but whether this can be overcome by altering the dose or duration of treatment remains to be seen.
Intravenous Albumin
Administration of intravenous albumin around the time of oocyte retrieval has been proposed as a measure to prevent OHSS, but the biological basis for this is unclear and the evidence of efficacy is poor. Two meta-analyses on this subject reached different conclusions [31, 54] and the evidence is characterised by methodological limitations and small subject numbers in individual trials. Further, the largest single-center trial on this subject did not find a protective effect of albumin administration in preventing OHSS [55].
Follicle Aspiration Prior to hCG Administration
It has been suggested that aspiration of some ovarian follicles prior to hCG administration may reduce the risk of OHSS by reducing the cohort of granulosa cells able to respond to HCG. The strategy involves an extra invasive procedure, which may make it less acceptable to patients than coasting. There is insufficient evidence to recommend this measure [46].
Choice of Luteal Support
hCG has a major role in precipitating OHSS and may worsen established OHSS. Progesterone is equally effective as hCG for luteal support and is associated with a lower risk of OHSS. It follows that hCG should not be used for luteal support. The increased risk of OHSS in multiple pregnancies [56] is further encouragement to adopt a policy of single embryo transfer in younger, more fertile women, who are also at a greater risk of OHSS. High-dose progesterone has been used in the luteal phase as a means of steroidal ovarian suppression, but there is no good evidence that this reduces the risk of OHSS beyond what would be achieved simply by the omission of hCG [57].
Management of OHSS
The most important aspect of the management of OHSS is for clinicians and patients to be aware of the condition. It is clear that OHSS cannot always be predicted when ovarian stimulation is being carried out. It follows from this that, in any woman receiving gonadotrophins, clinicians must remain conscious of the possibility of OHSS. Women undertaking such treatment should be given verbal and written information about OHSS including risk factors, symptoms, and signs of OHSS, what action to take and a 24-h contact number with prompt access to appropriately skilled clinicians.
Women with OHSS often present out-of-hours in acute hospitals separate from the assisted conception units in which they had their fertility treatment. It is important that fertility clinics establish close links with the acute units in their catchment area. Clinicians in acute hospitals may not be familiar with OHSS and should have access to evidence-based protocols and expert advice to aid their management.
Outpatient Management
Outpatient management is suitable for most cases of mild and moderate OHSS and, in certain circumstances, selected cases of severe OHSS. These patients mainly require supportive therapy and monitoring for progression of severity or onset of complications. Patients reporting symptoms suggestive of OHSS should be seen in order to make a diagnosis and assess the severity of OHSS. Clinical assessment should include observations, abdominal examination, respiratory and cardiovascular system review, body weight, and abdominal girth. Ultrasound examination to measure ovarian size and determine the presence or otherwise of abdominal free fluid is required to assess the severity of OHSS. Recommended laboratory investigations include haemoglobin, haematocrit, liver function tests, serum creatinine, and electrolytes. In patients reporting shortness of breath or those with severe OHSS, chest X-ray, chest ultrasound, and electrocardiogram may be helpful.
Abdominal discomfort can usually be managed with oral analgesia using paracetamol and opiates. Nonsteroidal anti-inflammatory agents should not be used, as they may compromise renal function. If the woman complains of severe pain, clinicians should take care to rule out alternative diagnoses, such as ovarian torsion, ectopic pregnancy, appendicitis, or pelvic infection.
Women with OHSS should be encouraged to drink to thirst, rather than to excess. Control of pain and relief of nausea may help maintain oral intake. Anti-emetics used should be compatible with the possibility of early pregnancy.
Women with OHSS are often advised to avoid sexual intercourse and strenuous exercise, as these may increase the risk of ovarian torsion. Light physical activity should be maintained, as reduced mobility may increase the risk of thrombosis.
In cases of mild and moderate OHSS, clinical review every 2–3 days is likely to be adequate; if the patient does not report any symptoms of deterioration, reviews may be conducted over the telephone. However, urgent clinical review in person is advisable if the woman experiences shortness of breath, increasing severity of pain, increasing abdominal distension, weight gain of ≥1 kg per day, or a perception of decreasing urine output.
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
