Comment on ‘Recombinant LH supplementation to a standard GnRH antagonist protocol in women of 35 years old or older undergoing IVF/ICSI: a randomized controlled multicentre study’

Sir,

I have enjoyed the lecture of the study recently published in your journal by König et al. (2013), where authors conclude that LH supplementation in GnRH antagonist cycles does not show any benefit in terms of pregnancy rate in patients 35 years old or more.

These findings could seem to be in discrepancy with those published by our group (Bosch et al., 2011). We observed that LH administration in antagonist cycles in patients aged 36–39 was associated with a significantly better implantation rate. Nevertheless, an analysis in detail of the differences between both studies draws interesting and complementary conclusions about the possible role of LH in the treatment of this particular population.

The methodological differences that may explain the inconsistency of the results are the use of a contraceptive pill (CP) the cycle prior to stimulation, and the substitution of 75 IU of recombinant (r) FSH by 75 IU of rLH from the first day of stimulation in the study group. These differences are reflected in the synthesis of estradiol (E2) and progesterone (P), and in the oocyte yield.

Although in our study hormonal determinations before starting stimulation were not available, it is very likely that after one cycle of CP all values were lower than in the present study. In this scenario, LH may help for a better steroidogenesis, promoting the synthesis of androgens as substrate for their later aromatization to estrogens. This is observed particularly in older patients, in whom basal androgens are decreased (Davison et al., 2005) and there is an impaired capability to produce androstenedione in response to rFSH (Welt et al., 2006).

The administration of rLH from the beginning of stimulation could be related to the lower P levels observed on the day of hCG. Through its action at the theca layer, LH enhances the conversion from pregnenolone to androstenedione, while FSH enhances its conversion into P in the granulosa cells. This P cannot be converted into androgens (Yding Andersen et al., 2011), so if its production is excessive it is delivered into the circulation (Fleming and Jenkins, 2010). In a multivariate analysis of more than 4000 cycles, we observed that P elevation is significantly related to the daily dose of FSH, but not of LH (Bosch et al., 2010).

The impact of LH on ovarian stimulation seems to be more patent when its administration is started at the beginning of the cycle than when given from the sixth day of stimulation, when follicular recruitment is already completed. In this case, only a modest increase in E2 and testosterone levels is observed, but probably too late to have an impact on ovarian response and cycle outcome. Indeed, no differences in follicular growth and oocyte yield are observed, while in our study, patients who received LH obtained less overall oocytes, but more metaphase II, reflecting a selective role of LH in ovarian response. This, together with lower P levels the day of hCG, may explain the better outcome of these patients when rLH is administered from stimulation Day 1.

In summary, the present study shows that LH supplementation from stimulation Day 6 in GnRH antagonist cycles does not improve cycle outcome in patients 35 years and older. This reinforces the concept that the possible beneficial effect of LH requires that its administration commences on the first day of stimulation to achieve optimal steroidogenesis and a better oocyte competence. This role may be especially needed when a CP is given on the previous cycle for programming. Further studies comparing both LH supplementation protocols may be necessary for an optimal address of this issue.