MERCK PUBLICATIONS FOR ESHRE - MINI PODCASTS

Ovarian stimulation with gonadotropins is an important part of IVF treatment, and IVF treatment is one of the cornerstones of modern management of infertility. The problem with gonadotropins is that they’re expensive.

Is there a way around that?
Possibly – by using biosimilars instead. They’re certainly cheaper.
The question is, do they work as well as the originator? And besides that, of course, are they as safe as the originator?

To try to find these two things out, we performed a systematic review and meta-analysis of clinical trial data. We looked at efficacy outcomes and risk of ovarian hyperstimulation syndrome, between biosimilars and originator recombinant follitropin alfa. This took us through the Medline, Embase, Cochrane, and Web of Science databases, from their origins all the way to January of 2019. To be thorough, we searched clinical trial registries as well, including ClinicalTrials.gov and the WHO international registry platform.

Our inclusion criteria were randomised, controlled trials that compared biosimilars with originator follitropin alfa. Our primary endpoint was live birth. Our secondary endpoints were clinical pregnancy and moderate to severe ovarian hyperstimulation syndrome. We used fixed effect meta-analysis to pool the data, and we calculated relative risks with 95% confidence intervals for each outcome. The I-squared statistic was our heterogeneity evaluation tool.

We found 370 candidate studies, identified 12 of them that were possibly useable by going through the abstracts, and finally found 3 that matched our inclusion criteria after going through the full publications. The biosimilar preparation in 2 of them was Bemfola, and Ovaleap was used in the other. More than 800 patients were included in those studies for GONAL-f and close to 950 patients for biosimilar preparations.

The result was interesting: Compared with the originator preparation, the live birth rate was significantly lower after the use of biosimilars (RR 0.82, 95% CI: 0.70 to 0.97), and so was the clinical pregnancy rate (RR 0.83, 95% CI: 0.71 to 0.96; 3 RCTs). The overall relative risk of hyperstimulation syndrome was non-significantly higher for biosimilars (RR 1.30; 95% CI: 0.83 to 2.70). Heterogeneity was negligible for all outcomes. (I-squared equalled zero percent.)

So it appears that, when looking at the current level of evidence given by clinical studies, biosimilar follitropin alfa preparations result in significantly lower live births and clinical pregnancy rates compared with the originator preparation. They don’t work as well, in other words, when looking at those clinical outcomes. For hyperstimulation risk, no difference was seen when comparing clinical trial results.

Our comparison study had its limitations. Meta-analysis was limited to the first treatment cycle, for one thing. Two studies did report the second cycle separately, but we couldn’t identify the participants with the same event in both cycles in the aggregated data. Also, women with histories of severe ovarian hyperstimulation syndrome were excluded from the trials.


Still, it’s useful to know that in published IVF trials so far, ovarian stimulation with biosimilar follitropin alfa reduced the chances of live birth, by 18%, to be exact, compared with the originator preparation.

For further information on GONAL-f, please refer to the prescribing information linked to this page on the podcast portal.

Recombinant human follicle-stimulating hormone (r-hFSH) is used for ovarian stimulation as part of fertility treatment, with individualized dosing (at start and during treatment) to optimize safety profile and efficacy. Owing to its mechanism of action there is a risk for OHSS with r-hFSH treatment, and there is also a risk for thromboembolic events, which increases in the presence of pregnancy and OHSS.

r-hFSH has been used for more than 18 years in clinical practice, and we wanted to find out what could be concluded about its safety profile from what’s reported in the literature and from Merck’s own safety data.

To that end, we systematically searched MEDLINE and Embase, from r-hFSH inception to the 19th of October, 2018, for clinical studies using it. Keywords were ‘GONAL-f’ and ‘r-hFSH’ and their variants. In a separate analysis, reports of OHSS and thromboembolism were obtained from the Merck KGaA Global Safety Database, covering November 29th, 2000 to October 19th, 2017. This database includes reports from healthcare professionals, patients, health authorities, clinical trials, non-interventional studies, and literature. Numbers of treatment cycles were estimated from sales data, based on an average 1875 IU administered per treatment cycle.

There were no exclusion criteria for the Global Safety Database search. The systematic review included clinical studies of patients with infertility receiving GONAL-f for ovulation induction or ART, with a starting dose within the range designated by the SmPC. Case reports and series were excluded, because they are included in the Global Safety Database. Data extracted were the number of patients exposed to GONAL-f, the number of treatment cycles, the incidence of OHSS, and the incidence of thromboembolism.

Results were reassuring. We’ve presented them separately for the systematic review and Global Safety Database search.

The systematic review identified 45 studies, comprising 5,186 patients exposed to GONAL-f and 5,240 treatment-cycles. Overall, we found 272 reported cases of OHSS (that’s 5,190 per 100,000 treatment cycles, or 5.19%) and 10 cases of severe OHSS (191 per 100,000 cycles, or 0.19%. There were no fatal cases. There were no reports of thromboembolism. The OHSS reporting rate may be high because of enhanced monitoring for adverse events in clinical trials. Also, patient characteristics in trials don’t fully reflect clinical practice.

In the Global Safety Database, there were 1,110 reported cases of OHSS (using the MedDRA preferred term ‘=OHSS’), and 80 reported cases of thromboembolic events. Overall, there were an estimated 16,525,975 treatment cycles since November 29th, 2000. This resulted in reporting rates for OHSS of 6.7 per 100,000 treatment cycles (or 0.007%) and 0.48 per 100,000 cycles (or 0.0005%) for thromboembolism. Three fatal cases were recorded: two of OHSS and one of thromboembolism. The OHSS frequency reported from the Safety Database may be lower than the systematic review because it reflects real-world reporting and includes more sources of safety data.

So what did we show with all this? Basically, OHSS and thromboembolism risk after ovarian stimulation with GONAL-f were shown to be low in our analysis.

It’s true that our study had its limitations. The cases reported in the Global Safety Database were either submitted voluntarily or obtained from the literature. There may therefore be some duplication of cases between the Global Safety Database and the systematic review. In addition, the number of treatment cycles in the post-marketing setting had to be estimated, from sales data.

Still, the wider implications of what we found are useful. GONAL-f has been in use for more than 18 years, with low rates of OHSS and thromboembolism, and that is reassuring.

For further information on GONAL-f, please refer to the prescribing information linked to this page on the podcast portal.

Eeva is a time-lapse embryo-monitoring system with integrated software that automatically analyses embryo images to define timings of early cell divisions and make a recommendation on the likelihood that an embryo will develop to the blastocyst stage. Eeva predicts blastocyst formation at cleavage stage with a specificity of 85%.

However, results from several trials of Eeva to date demonstrate substantial heterogeneity (e.g., different embryo rankings, different primary endpoints, or lack of generalizability across different centres). This means that the true benefit of Eeva as an adjunct to morphological assessment for embryo prediction has not been determined.

It would be useful to know whether implantation rates for Day 3/5 embryo transfers selected by early embryo viability assessment (that’s Eeva) plus morphology grading (we’ll call that ‘MG’) are different from transfers selected only by MG.

We conducted a Phase IV, prospective, 2:1 randomized, exploratory clinical trial at 20 sites in eight countries (Canada, Germany, Italy, France, Norway, Sweden, Spain and UK). 970 participants were randomly assigned to Eeva+MG or only MG. 899 completed the trial: that was 593 (66%) in the Eeva+MG group and 306 (34%) in the MG group. The choice of embryo for transfer was determined by the embryologist.

Women diagnosed with infertility who had failed 3 or fewer IVF/ICSI cycles were included. Participant recruitment continued for 9 months, and participants were followed up to Days 12 through 18 to verify implantation and up to gestational Weeks 5 through 8 to verify clinical pregnancy. The study duration was 19 months. The original primary endpoint (positive implantation with fetal heartbeat) was amended to “implantation with intrauterine gestational sac” after database lock, owing to errors when recording the original endpoint.

Results? The implantation rate with intrauterine gestational sac at Week 8 was 30.1% in the Eeva+MG group and 35.3% in the MG-only group. Post-hoc supportive analyses showed that only 57.7% of patients in the trial met both compliance criteria, implying that almost a half of patients in the Eeva+MG group were assessed on MG alone or had incorrect application of Eeva. Additional post-hoc supportive subgroup analyses of the original endpoint in the Eeva-compliant subgroup showed the implantation rate with fetal heartbeat after Day 3 embryo transfer to be 25.7% in the Eeva+MG group (there were 163 of them) and 32.7% in the MG group (the ‘n’ was 117). For Day 5/6 transfer, the implantation rate was 45.5% in the Eeva+MG group (n=87) and 38.3% in the MG group (n=99). The implantation rate of the Eeva-compliant subgroup was equal to, or higher than, that of the MG group at five of 12 sites (Day 3 transfer) and seven of 14 sites (Day 5/6 transfer).

We can summarize, therefore, that implantation rate in our study did not differ between Eeva+MG and MG alone, which was possibly due to the high rate of non-compliance in the Eeva+MG group.

Our study had its limitations. The low compliance was based on reliance on MG scores over Eeva scores and the suboptimal use of Eeva, possibly because of insufficient training of embryologists – which means that the true value of Eeva may not have been fully leveraged by these results.

Therefore, we need to conclude that the results of our trial are in line with the (still inconclusive) results reported in other trials of Eeva. Our findings also highlight potential limitations in the conduct of trials of devices that are supposed to improve embryo assessment, and they also point to likely gaps in the training of embryologists in how to use the Eeva system.

Cetrorelix, Ganirelix and Teverelix are all GnRH antagonists with decapeptidic structures that differ by only one or two amino acids. Cetrorelix and Ganirelix are used in ovarian stimulation, to help control luteinizing hormone production. Some reports have suggested that these antagonists have different potencies at the cellular level, which could lead to different outcomes. But these differences are not well understood.

So, we designed an in-vitro model study of different doses of Cetrorelix, Ganirelix and Teverelix, ranging from picomoles to micromoles, and had a look at GnRH responses in the short- and long-term – that is, for up to half an hour, and up to 24 hours. We compared 3 to 6 treated samples, and we used untreated samples as controls.

The cells we used were GnRH receptor I-transfected HEK293 and neuroblastoma SH-SY5Y. We also used the LβT2 mouse pituitary cell lines, which express the GnRH receptor. We evaluated short-term and long-term inhibition of GnRH-induction by different methodologies, including bioluminescence, western blotting, immunostaining and Real -time PCR, and followed as intracellular signaling factors cAMP, pCREB, pERK1/2 and β-catenin activation, as well as Ca2+ increase and Lhb gene expression, by real-time PCR.

Our results? In transfected HEK293 cells, 10 nanomoles of Cetrorelix inhibited GnRH-induced intracellular Ca2+ increase, whereas similar inhibition was seen with Ganirelix and Teverelix at a ten-fold higher concentration. These findings were confirmed in GnRHR-transfected SH-SY5Y cells, where the 3 x 50% effective concentrations of GnRH antagonists were 10-fold higher than in HEK293 cells, probably because of differences in receptor expression at the cell surface between the two transfected cell lines. Untransfected HEK293 and SH-SY5Y cells did not reveal any detectable response to GnRH. Cetrorelix had higher efficacy than Ganirelix or Teverelix, resulting in an inhibitory concentration of 1.56±2.49 nM, which was lower than the inhibitory concentrations of Ganirelix (16.60±3.76) or Teverelix (62.80±3.77)

In spite of the differences among these drugs in inhibiting GnRH mediated short-term cell response, the antagonists depleted pCREB, pERK1/2 and β-catenin activation at similar concentrations. (The range was 100 nM–1.0 µM.) Also, all the antagonists tested inhibited GnRH-induced Lhb expression in LẞT2 cells, which suggests that they may efficiently target luteinizing hormone production.

Our study had its limitations. This was in-vitro, and the differences we saw in potency or efficacy for the three antagonists might not translate in-vivo, since the metabolic clearance and pharmacokinetics may differ among them, especially if they’re in differing formulations and solubility.

But it is good to know that the GnRH antagonists Cetrorelix, Ganirelix and Teverelix are different with respect to in-vitro activity for inducing intracellular Ca2+ increase. This by itself may lead to different short-term regulation in the target cells, and that could well modulate in-vivo outcomes.

For further information on Cetrorelix, please refer to the prescribing information linked to this page on the podcast portal.