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

bulk of the existing data regarding the use of luteal estradiol priming. Given our use of observational data we made every effort to minimize the faults of the individual studies included in our manuscript, using rigorous methodology and the random effects model to minimize the inter-study differences and improve the applicability of our results to all populations of poor responders across the spectrum from mild to severe. It is because of this heterogeneity that we chose a random effects model (DerSimonian and Laird, 1986) to minimize the intrinsic effects of population variation and increase the generalizability of our results. In doing so we believe that the results of our analysis can be applied to patients on both ends of the spectrum—patients with only a mild degree of poor ovarian response and those with a response so consistently poor that they are repeatedly cancelled prior to retrieval. By incorporating studies that included both mild and severe ends of the poor ovarian response spectrum as highlighted by Drs Polyzos and Tournaye, this strengthens our findings by improving the applicability of our results instead of isolating the findings to a more severe phenotype. Drs Polyzos and Tournaye question the effect of publication bias on our findings. We appreciate this comment, and we do agree that there may be some degree of publication bias included in our analysis, as could be found in any meta-analysis. Our systematic review included studies with both positive and negative findings, and we incorporated all of these results into our analysis. In truth, one of our primary goals in performing this analysis was to improve the body of literature evaluating treatment of the poor responder, and if the results of our analysis lead to increased publications describing a neutral or negative effect of the luteal estradiol protocol in poor responders, this achieves our goal in improving available data which we welcome greatly. As with any study, we do agree with Drs Polyzos and Tournaye that our findings should be interpreted with caution as cited in our paper as these findings are limited by the body of literature currently available. As the poor responder lacks a concrete definition, there is some heterogeneity to these results, which merits caution when applying our findings to individual patients. Furthermore, the increased clinical pregnancy rate demonstrated when using the LE protocol may be principally related to the decreased cycle cancellation rate. Despite these limitations, …