The prevention of ovarian hyperstimulation syndrome

Ovarian hyperstimulation syndrome (OHSS) is a serious, potentially life-threatening complication of certain infertility treatments. It may occur following attempts to induce unifollicular ovulation in anovulatory infertile patients, but principally OHSS follows controlled ovarian hyperstimulation to increase the number of oocytes available for assisted conception. The incidence of significant OHSS complicating assisted conception is variably quoted as between 0.6% to 14%'. Given the iatrogenic nature of this condition, it is particularly important to minimise the risk of OHSS in women receiving fertility treatments. This review considers how this may be achieved. An understanding of the pathophysiology of OHSS is essential to the rational development of strategies for preventing it. The underlying cause of OHSS is unknown, but a vasoactive ovarian factor is likely to be involved. The syndrome develops following supraphysiological ovarian stimulation and requires the trigger of either human chorionic gonadotrophin (hCG) or luteinising hormone (LH). It is recognised that OHSS is more common in conception treatment cycles than nonconception treatment cycles and in cycles where hCG rather than progesterone is used for luteal support2. It may be more frequent and more severe in cases of multiple pregnancy3, where hCG levels are higher than in singletons from an early stage of gestation4. A number of studies, including some using animal models, have suggested various pathophysiological mechanisms for the development of OHSS, which have formed the basis of preventative measures for OHSS, targeting specific mechanism^^-^. A distinction should however be made between primary pathophysiological processes and compensatory mechanisms that are secondarily activated in the course of OHSS'. Many strategies for the prevention of OHSS rely on the ability to predict the probability of its occurrence in the individual patient. Ideally patients at risk of developing OHSS should be identified prior to treatment and epidemiological studies have attempted to do this. A retrospective multicentre study in Belgium'' examined the clinical features of 128 patients with OHSS and reported an association with young age and endocrinopathy. Women with polycystic ovarian syndrome (PCOS) have been shown to be at particular risk of OHSS"*12. Even when PCOS is not suggested by endocrine results, OHSS is associated with the necklace sign on ultrasound scan, in which the ovary in the luteal phase shows sonographic features similar to PCOS of small peripherally placed follicles with an echodense ~troma'O*'~. A link between lean habitus and OHSS has been demonstrated by some14, but others have not found this predictive15. It is not possible to predict reliably all cases of OHSS from pretreatment patient characterist i c ~ ' ~ , 16 thus attention has focused on identifying risk factors that are manifest during treatment. The sensitivity of the ovary to gonadotrophic stimulation and the magnitude of the ovarian response to gonadotrophins appear to be important determinants of the probability of developing OHSS. Women who develop OHSS require significantly lower doses of hMG for stimulation and have a lower ratio of hMG dose to duration of stimulation than matched controls13* 16. It has been suggested that the rate of rise of serum oestradiol is more important than the absolute levels, with very rapid rises reflecting hypersensitivity of the ovary to stimulation16. l ' . Peak levels of oestradiol are significantly higher in in vitro fertilisation (IVF) cycles complicated by OHSS than in control It was anticipated that it would be possible to define an upper limit of serum oestradiol, above which the risk of OHSS would be so high that the treatment cycle should be abandoned, thereby avoiding OHSS. However, there is a lack of agreement as to the upper limit of oestradiol which is felt to constitute a significant risk of OHSS, with different investigators quoting values from 1500 pg/ml (5505 pmol/L) to greater than 6000 pg/ml (22,020 pmol /L)"~~~17,18. Levels of oestradiol alone do not provide a clinically useful guide to the risk of OHSS, both because of the considerable overlap with normal cycles and the variation in assay results between laboratories. Further, in a