**3.1. Summary points**

60 Enhancing Success of Assisted Reproduction

from cycles utilizing a GnRH agonist protocols.

Author, Year RTCs Number

of Patients

Absolute Pregnancy Benefit

Al-Inany, 2005 8 2031 - - 1.18 (0.93-1.50) Al-Inany, 2008 11 2937 +3% +21% 1.20 (1.01-1.42) Coomarasamy, 2008 7 2159 +4% +18% 1.18 (1.02-1.38) Jee, 2010 5 2299 +3% +12% 1.14 (0.98-1.33) Lehert, 2010 16 3952 +3% +10% 1.10 (0.97-1.25)

**Table 1.** Recent meta-analysis comparing hMG versus rFSH for ovarian stimulation in ART cycles.

A recent RCT published in 2012 has provided similar evidence for the benefit of hMG in GnRH antagonist cycles. Devroey *et al.* randomized 749 patients to receive either hMG or rFSH (79). There were numerous strengths to this trial: rigorously described randomization, allocation and concealment, the use of 25 clinics in 7 countries, all patients were only allowed a single blastocyst transfer, and the follow-up included live births from the fresh cycle plus subsequent frozen cycles of embryos obtained during the study. Patients in the hMG arm had higher estradiol, LH, and FSH measured in the serum on the day of hCG. There was a significant reduction in the number of oocytes retrieved in the hMG group (-1.6 oocytes per retrieval). Importantly, an absolute difference of +3% in live birth with the use of

Relative Pregnancy Benefit

Pregnancy OR (95% CI)

There is a large body of randomized controlled trials available for analysis comparing hMG to rFSH only. These trials are relatively homogenous, with similar dosing strategies and primarily GnRH agonist pituitary downregulation. These RCT have been systematically evaluated in several meta-analyses shown in Table 1 (74-78). The number of patients required to show a benefit in hMG had been calculated at over 2100 (76). This is demonstrated in a 2005 meta-analysis by Al-Inany *et al.* where 8 RCTs including 2031 ART cycles failed to show a statistically significant improvement in live birth (OR 1.18, 95%CI 0.93-1.50) although a trend to benefit may have been seen (76). When the same authors repeated a meta-analysis in 2008, there were 11 RCTs including over 2900 patients available for analysis (75). This time a significant improvement in live birth (OR 1.20, 95%CI 1.01-1.42) was demonstrated with the use of hMG versus rFSH alone (75). This data was confirmed in a separate meta-analysis by Coomarasamy *et al*. showing an improvement in live birth with hMG (OR 1.18, 95%CI 1.02-1.38) (77). Two more recent meta-analyses in 2010 each failed to show a significant improvement in live birth with the use of hMG (74, 78). However, the p values for these studies were borderline significant (0.051-0.06) and the odds ratio of pregnancy was similar to the other trials. Indeed, the last four meta-analyses had all demonstrated between a 3-4% absolute increase in pregnancy and a 10-21% relative increase in pregnancy with the use of hMG as compared to rFSH alone. These numbers translate to a NNT of approximately 32 patients with hMG to achieve one additional live birth. The clinical relevance of this number has been a matter of debate, but there is a clear statistical benefit to utilizing hMG. The majority of these source RCTs for these meta-analysis were

