Addition of neither recombinant nor urinary luteinizing hormone was associated with an improvement in the outcome of autologous in vitro fertilization/intracytoplasmatic sperm injection cycles under regular clinical settings...

We reviewed 22,212 in vitro fertilization/intracytoplasmic sperm injection cycles and found that the addition of luteinizing hormone was not associated with an improvement in the outcome.

Volume 106, Issue 7, Pages 1714-1717


Juan-Enrique Schwarze, M.D., M.Sc., Javier A. Crosby, Ph.D., Fernando Zegers-Hochschild, M.D.



To determine whether the addition of either urinary or recombinant LH in patients undergoing routine clinical care improved the outcome in terms of the number of oocytes recovered for insemination or the delivery rate per initiated cycle.


Cohort analysis.


Couples undergoing IVF/ICSI in 158 institutions in 15 countries in Latin America.


In vitro fertilization clinics.


We compared the outcome of three different protocols of COH, including rFSH only, rFSH plus rLH, and rFSH plus hMG.

Main Outcome Measure(s)

The number of mature oocytes recovered and inseminated; proportion of ETs at the blastocyst stage; clinical pregnancy, miscarriage, and delivery rates; proportion of cycles with embryo cryopreservation; and mean number of embryos cryopreserved.


After correcting for the age of the female partner, body mass index, number of embryos transferred, and stage of embryo development at transfer, we found that LH addition was not associated with an increase in the mean number of metaphase II oocytes inseminated or with an increase in the delivery rate or changes in the miscarriage rate.


Our study strongly suggests that in routine clinical practice, the type of controlled ovarian stimulation—FSH alone or in combination with LH—has little impact on the outcome of assisted reproductive technology; therefore a more friendly and accessible alternative should be favored.

Read the full text here.


Go to the profile of Micah J Hill
almost 6 years ago
Dear Editors, I read the article from Schwarze et al. with great interest. They conclude that the addition of rLH or hMG to IVF ovarian stimulation has little impact on outcomes and that more friendly and accessible alternatives should be favored. For several reasons I disagree with their conclusions. First, the analysis of live birth was done with two different units of measurements, both inappropriate. Live birth was analyzed on a per oocyte retrieval and per embryo transfer basis. However, the intervention and effect of gonadotropin stimulation occurs well before this. The appropriate outcome should be live birth per cycle start. Any analysis other than per cycle start has the potential for bias and loss of valuable information. Patients who had all embryos frozen without fresh embryo transfer and those with poor response and no retrieval or transfer would be lost in this present analysis. This unit of analysis issue introduces the potential risk for substantial bias (1). Second, there is clearly inherent potential bias in the retrospective design of a study conducted across 158 IVF practices in 15 different countries. It is incumbent in such study designs to make maximal effort to control for potential confounding variables. IVF practices across 15 countries will vary in what gonadotropins are available and utilized locally. These practices will also vary in their patient populations, IVF protocols, and laboratory conditions. Live birth rates across 158 practices will vary widely for a myriad of reasons, most of which are likely to be unrelated to gonadotropin protocols. If a practice with a good laboratory and high live birth rate utilizes primarily rFSH only protocols, this would clearly bias the data in one direction, potentially unrelated to the gonadotropin itself. This is why even RCTs performed in multiple centers will attempt to account for this effect by the modeling of each center into the statistical analysis (2). Unfortunately the present study did not. There is a wide range of other potential confounding variables in this present study, but the authors account for only 4 variables in the logistic regression model. Other confounding factors such as patient diagnosis were not accounted for. For example, there was a diagnosis of ovarian insufficiency in 6.7% of cases using rFSH only, versus 11.7% and 16% in the hMG and rLH groups. Surprisingly, the authors do not tell us any statistical comparisons of the baseline differences between the groups of their retrospective study. Further, they don’t account for these differences in the regression modeling, leaving us to wonder if the outcomes are biases by unaccounted confounding variables. Indeed, taking their data, the P value is <0.0001 for the comparison of the diagnosis of ovarian insufficiency between the groups, clearly biasing the data in favor of rFSH only protocols. Unfortunately the authors do not tell us that baseline differences exist, nor do they account for them in their analysis. Finally, the authors cite the most recent Cochrane review on the topic in favor of their conclusions (3). However, the Cochrane review concluded that there are indeed small differences in effectiveness between gonadotropins. In 11 randomized controlled trials totaling 3197 subjects, rFSH only protocols decrease live birth (OR 0.84, CI 0.72-0.99) (3). Other meta-analyses demonstrate the same finding, with an absolute reduction in live birth of between 3-4% with rFSH only as compared to rFSH + hMG (4,5). Yet the authors of the current study do not cite this information. The question then becomes how does a retrospective study, with its inherent unaccounted for biases, inform the hypothesis in question more than 11 randomized controlled trials? Does this study’s opposite conclusion compared to level I data change our opinion on the subject? A 3% difference in live birth may or may not be clinically relevant based on the costs and the risks of the intervention. The NNT for 1 additional live birth with the addition of hMG to IVF protocols is 33. In my opinion, this is a reasonable number to treat, given the similarity in cost and minimal risks associated with adding hMG to a stimulation protocol. Others may reasonably disagree. But let us judge the clinical question on the absolute risk difference and NNT from an abundance of level 1 data, not a single retrospective study. Sincerely, Micah J Hill 1) Wilkinson J, Roberts SA, Showell M, Brison DR, Vail A. No common denominator: a review of outcome measures in IVF RCTS. Hum Reprod in press 2) The Corifollitropin Alfa Dose-finding Study Group. A randomized dose-response trial of a single injection of corifollitropin alfa to sustain multifollicular growth during controlled ovarian stimulation. Hum Reprod 2008;23:2484-92. 3) Van Wely M, Kwan I, Burl AL, Thomas J, Vail A, Van der Veen F, Al-Inany HG. Recombinant versus urinary gonadotrophin for ovarian stimulation in assisted reproductive technology cycles. A Cochrane review. Hum Reprod Update 2012:18:111 4) Al-Inany HG, Abou-Setta AM, Aboulghar MA, Mansour RT, Serour GI. Efficacy and safety of human menopausal gonadotrophins versus meta-analysis recombinant FSH: a mata-analysis. Reprod Biomed Online. 2008;16:81-8. 5) Coomarasamy A, Afnan M, Cheema D, van der Veen F, Bossuyt PMM, van Wely M. Urinary hMG versus recombinant FSH for controlled ovarian hyperstimulation following an agonist long down-regulation protocol in IVF or ICSI treatment: a systematic review and meta-analysis. Hum Reprod 2008;23:310-5.