Approximately 15% of women with PCOS remain anovulatory despite treatment with oral anti-oestrogen medications such as clomiphene citrate. In addition, about half of women with PCOS ovulating on anti-oestrogen treatment fail to conceive. Gonadotrophin stimulation is the next step in treatment for women who are "clomiphene resistant", however, results of gonadotrophin stimulation in women with PCOS are less successful. In PCOS associated with hypersecretion of LH, purified urinary follicle-stimulating hormone (u-FSH) preparations have theoretical advantages over the use of human menopausal gonadotrophin (hMG) preparations (containing both FSH and LH) , but whether this claimed advantage extends into clinical practice remains uncertain. In addition, the use of gonadotrophin-releasing hormone analogues (GnRH-a) to produce pituitary desensitisation prior to ovulation induction in PCOS has been claimed to increase the success rates of treatment as well as reduce complications such as OHSS and multiple pregnancy. Gonadotrophin preparations have also been administered via different routes (intramuscular or subcutaneous), or using different stimulation regimens and protocols (step-up or standard) in an attempt to improve efficacy.
To determine the effectiveness of urinary-derived gonadotrophins as ovulation induction agents in patients with PCOS trying to conceive. In particular, to assess the effectiveness of (1) different gonadotrophin preparations, (2) the addition of a gonadotrophin-releasing hormone agonist (GnRH-a) to gonadotrophin stimulation and (3) different modalities of gonadotrophin administration.
The search strategy to identify RCTs consisted of (1) the Group's Specialised Register of Controlled Trials using the search strategy developed for the Menstrual Disorders and Subfertility Group as a whole (see the Review Group details for more information), (2) additional specific electronic MEDLINE searches and (3) bibliographies of identified studies and narrative reviews.
RCTs in which urinary-derived gonadotrophins were used for ovulation induction in patients with primary or secondary subfertility attributable to PCOS.
Data collection and analysis
Twenty three RCTs were identified, nine of which were excluded from analysis. The data were extracted independently by two authors. The following criteria were assessed:
(1) the methodological characteristics of the trials
(2) the baseline characteristics of the studied groups and
(3) the outcomes of interest: pregnancy rate (per cycle), ovulation rate (per cycle), miscarriage rate (per pregnancy), multiple pregnancy rate (per pregnancy), overstimulation rate (per cycle) and ovarian hyperstimulation syndrome (OHSS) rate (per cycle). Where suitable, meta-analysis was performed using Peto's OR with 95% CI with the fixed effect Mantel-Haentszel equation.
(1) A reduction in the incidence of OHSS with FSH compared to hMG in stimulation cycles without the concomitant use of a GnRH-a (OR 0.20; 95% CI 0.08-0.46) and
(2) a higher overstimulation rate when a GnRH-a is added to gonadotrophins (OR 3.15; 95% CI 1.48-6.70).
Although 14 RCTs were included in this review, few dealt with the same comparisons, all were small to moderate size and their methodological quality was generally poor. Any conclusions, therefore, remain tentative as they are based on a limited amount of data and will require further RCTs to substantiate them. In none of the comparisons was there a significant improvement in pregnancy rate but this may be due to the lack of power (i.e. insufficient patients randomised to demonstrate a significant difference between treatments). There was a trend towards better pregnancy rates with the addition of a GnRH-a to gonadotrophin stimulation and these interventions warrant further study. Despite theoretical advantages, urinary-derived FSH preparations did not improve pregnancy rates when compared to traditional and cheaper hMG preparations; their only demonstrable benefit was a reduced risk of OHSS in cycles when administered without the concomitant use of a GnRH-a. No conclusions can be drawn on miscarriage and multiple pregnancy rates due to insufficient reporting of these outcomes in the trials.