A comparison of menotropin, highly-purified menotropin and follitropin alfa in cycles of intracytoplasmic sperm injection

Over the last several decades, a marked development has been made in the field of gonadotropins for ovarian stimulation. This development is the result of an evolution in manufacturing, with the production of therapeutic gonadotropins changing from a simple process of urine extraction and purification to a complex system involving recombinant technology, yielding high levels of purity, quality, and consistency. Menotropin (human menopausal gonadotropin [hMG]), labeled as a 1:1 ratio of follicle-stimulating hormone (FSH) and luteinizing hormone (LH) activity, has been widely used since the 1960s. As the purification processes became more sophisticated, other purified gonadotropin products were introduced, including a menotropin formulation (highly-purified hMG; HP-hMG) with the same labeled ratio of FSH:LH activity. In the 1980s, biotechnology advances resulted in the formulation of a recombinant human FSH (r-hFSH), follitropin alfa, the first r-hFSH to become available in the market. Another r-hFSH product, follitropin beta, was also launched [1]. Due primarily to its safety, efficacy, and ease-of-use, r-hFSH soon became the preferred gonadotropin worldwide for controlled ovarian hyperstimulation (COH) in the context of in vitro fertilization (IVF) cycles.

In current practice, factors such as efficacy, safety, cost of the medication, ease of administration, and availability of product are the determinants of gonadotropin choice. However, in Brazil and other Latin American countries, assisted reproductive techniques (ART) and the drugs used for ovarian stimulation are neither provided by healthcare plans nor reimbursed by the government. Thus, patients prefer to purchase lower-cost drugs. In recent years, our choice of gonadotropin product was motivated by patient costs and, therefore, we justified the choice of urinary menotropins for COH in ART procedures. When HP-hMG became available, we recommended this product to our patients. However, in addition to safety, efficacy, and purity, follitropin alfa, an r-hFSH product, has a highly consistent isoform profile and low batch-to-batch variability that may be associated with a more predictable follicular development [2]. Therefore, in 2004, we selected r-hFSH as the drug of choice for COH in intracytoplasmic sperm injection (ICSI) cycles.

The objective of the current analysis was to assess treatment outcomes of ART procedures at our center to compare the efficacy of hMG, HP-hMG, and r-hFSH (follitropin alfa filled-by-mass [FbM]) for COH after pituitary down-regulation in ICSI cycles.

Consecutive ICSI cycles (n = 865) performed at Androfert, a tertiary referral center for male infertility treatment in Campinas, Brazil, from April 2000 to November 2005 were included in this analysis. Laboratory and clinical protocols remained unchanged during these time periods, except for the type of gonadotropin that was sequentially changed over time, as follows: hMG (Menogon^®, Ferring, Brazil) from April 2000 to February 2002; HP-hMG (Menopur^®, Ferring, Brazil) from March 2002 to January 2004; and r-hFSH (GONAL-f^® FbM, Merck Serono, Brazil) from February 2004 to present. During the considered period, COH was performed in 299 cycles with hMG; in 330 cycles with HP-hMG; and in 236 cycles with r-hFSH.

Patient cycle data were entered into a database using customized software (Androsys^®, Androfert-UNIOESTE, Brazil). The overall outcomes of our IVF program have been consistent during the study period (Table 1).

Signed informed consent was obtained from patients to use both clinical and laboratory data for analysis. The analysis was approved by our Institutional Review Board. Indications for ICSI were in accordance with the guidelines of the II Brazilian Consensus of Male Infertility [3], even if the indication for IVF was a female factor.

The distribution of the 865 ICSI cycles (age, indication, baseline serum FSH and LH levels, and total number of ICSI cycles) is presented in Table 2. The mean ± standard deviation (SD) patient age was significantly lower in the hMG group (32.5 ± 5.3 years) compared with patients in the HP-hMG and r-hFSH groups (34.0 ± 4.7 and 34.3 ± 4.8; p < 0.001, respectively). Even after stratification by patient age groups, 71.2% of patients were ≤35 years old in the hMG group (mean ± SD age = 29.9 ± 3.7 years), compared with 60.0% (mean ± SD age = 30.9 ± 3.2 years) and 60.2% (mean ± SD age = 30.9 ± 3.2 years; p = 0.005) of the HP-hMG and r-hFSH groups, respectively. In the >35 years old patient subgroup, the mean ± SD baseline LH levels were significantly lower in the hMG and r-hFSH groups (4.2 ± 3.0 and 4.3 ± 3.9, respectively) than in the HP-hMG group (5.8 ± 4.8; p = 0.004). Baseline FSH was not significantly different between groups.

The clinical and laboratory outcomes are provided in detail in Table 3. The length of stimulation was significantly shorter in the hMG group in comparison with the other groups, both for the total group and the age-stratified subgroups. On the day of hCG administration, the serum estradiol levels were significantly different between the different gonadotropin product groups and for the age-stratified subgroups as well. On the other hand, the number of metaphase II oocytes in patients >35 years of age treated with hMG was significantly lower when compared with the other treatment groups. The percentage of top quality embryos available for transfer was significantly higher in the HP-hMG group, both for the total group and the subgroup ≤35 years of age. The cancellation rates due to poor ovarian response and the incidence of serious OHSS did not differ among the groups.

The clinical pregnancy rates per initiated cycle, implantation and live birth rates per initiated cycle were not statistically different among the groups (Table 4). However, the differences seen in the live birth rates were in favor of the r-hFSH and HP-hMG groups, for both the total population (p = 0.09) and the subgroup of patients ≤35 years of age (p = 0.08; Table 4), although these differences were not statistically significant. After stratification for indication and number of ICSI attempts, there was no difference in the live birth rates among the groups (data not shown, p = 0.16). The incidence of spontaneous abortion was significantly higher in the hMG group compared with the other groups (p = 0.009), mainly due to the subgroup of patients >35 years of age (Table 4).

The clinical and laboratory outcomes of cycles with clinical pregnancy are provided in Table 5. The mean ± SD patient age and the length of stimulation were significantly lower in the hMG group compared with the other groups, both for the total group (p = 0.01) and for the age-stratified subgroup ≤35 years (p = 0.004 and p = 0.001, respectively). On the day of hCG administration, serum estradiol was significantly different between the hMG and HP-hMG groups (p = 0.005) and for all gonadotropin products in the ≤35 years subgroup as well (p = 0.01). The percentage of top quality embryos available for transfer on day 3 was significantly higher in the HP-hMG group, both for the total group (p = 0.004) and the ≤35 years subgroup (p = 0.02). Baseline FSH and LH levels, mean number of retrieved and mature oocytes, fertilization rate, and the mean number of transferred embryos were not significantly different between groups.

The distribution of live birth cycles in relation to parity, gestational age, and birth weight are shown in Table 6. Singleton births were seen in 68.5%, 63.5%, and 64.8% of live birth cycles in the hMG, HP-hMG and r-hFSH groups, respectively. Significant differences were observed for the gestational age, in the three parity stratifications, in favor of the r-hFSH group compared with the hMG group. In the case of twins and triplets, the difference was significant versus the r-hFSH group (p = 0.03). For triplets, the birth weight was significantly higher in the r-hFSH group versus the hMG group (p = 0.03).

The gonadotropin dose used per cycle was significantly lower in the r-hFSH group compared with the other groups, for both the overall population (p < 0.001) and the stratified subgroups (p < 0.001 and p = 0.02 for the ≤35 years and >35 years subgroups, respectively) (Table 7). These differences were also significant for r-hFSH after stratifying for the cause of infertility (male and mixed factors) and for the number of treatment cycles. Taking into account only the cycles that resulted in clinical pregnancy and live birth, the mean total doses of gonadotropins used for ovarian stimulation were significantly lower for the r-hFSH group in comparison with the hMG and HP-hMG groups.

Gonadotropin dose adjustments were performed more often in the r-hFSH group (53.4%) than in the hMG (18.7%) and HP-hMG (20.3%) groups (p < 0.001; Table 3). The total dose of gonadotropin to achieve a pregnancy with a live birth was significantly higher in the hMG and HP-hMG compared with the r-hFSH group (p = 0.02; Table 8). The necessary dose of hMG to achieve a pregnancy was 15.7% higher than the r-hFSH dose, while the necessary dose of HP-hMG was 11% higher. These differences were still higher when considering the live birth parameter (45.3% and 19.8%, for hMG and HP-hMG, respectively) (Table 9). Because relative differences obtained from the rates cannot be statistically compared, it was not possible to calculate whether these differences were significant.

IVF is a highly complex procedure. Before clinical pregnancy and live birth can be achieved, several critical steps are required, including the ovarian stimulation regimen. In this context, it is vital to evaluate these outcomes in routine clinical practice to account for subtle differences in patient response. As a rule, comparative, prospective and randomized clinical trials are the gold standard to evaluate the efficacy and safety of a therapy. However, some of these trials are not statistically powered to show relevant clinical differences between treatments. Moreover, strict inclusion and exclusion criteria are required, which do not represent the heterogeneous patient population typically encountered in clinical practice. Another important consideration is that many ART studies use the clinical pregnancy rate as the primary efficacy endpoint instead of live births. However, the latter may be difficult to evaluate if pregnancy outcome data are not readily available.

The current analysis examined the clinical efficacy of three different gonadotropins used for COH in ICSI cycles. In our program, clinical and laboratory data are systematically and continually entered into our patient database and managed by software specially developed for assisted reproduction programs (Androsys^®, Brazil). Clinical follow up of pregnancies is carried out twice per month. With confirmation of live birth, follow up continues with the registration of gestational ages, birth weights, neonatal disorders, and eventual malformations for a period of 30 days post delivery. Using the Androsys^® database, it was possible to compare COH regimens in more than 860 consecutive ICSI cycles at a single ART center. To our knowledge, this is one of the most comprehensive comparative analyses examining both the live birth rates and the total gonadotropin dose needed per cycle resulting in clinical pregnancy and live birth. Additionally, our COH regimens consisted of the respective gonadotropin products (hMG, HP-hMG, and r-hFSH) administered as single agents, that is, combination protocols of hMG or HP-hMG plus r-hFSH were not used, a clinical approach that is widely utilized elsewhere.

Due to the retrospective, observational design of this analysis, the possibility of some inherent bias exists, although the non-selected patient population was representative of the therapeutic profile observed in current clinical practice. Outcomes were analyzed on an overall basis and stratified according to acknowledged prognostic parameters commonly used in reproductive medicine (age, infertility factor, and number of treatment attempts) to account for potential bias. For example, demographic characteristics were not homogeneous across the patient population. In fact, the mean age of the menotropin (hMG) group was significantly lower (1.5 years on average) and baseline LH values were significantly higher in the HP-hMG group than in the other groups. Moreover, more than 70% of the patients treated in the hMG group were aged ≤35 years versus approximately 60% in the other two groups. This could explain the shorter duration of stimulation observed in this group as compared with the HP-hMG and r-hFSH groups.

The three treatment groups showed similar live birth rates independent of the gonadotropin used for COH, although a trend towards higher live birth rates was observed in the r-hFSH and hMG-HP groups, in both the overall population and the subgroup aged ≤35 years. In 2004, Ludwig et al. published their analysis of more than 20,000 cycles from the IVF National Registers in Germany and found that live birth rates were 16.6% higher with r-hFSH compared with hMG when used for COH after pituitary down-regulation [10]. In their analysis, however, the authors did not evaluate the results stratified by the type of hMG (conventional or purified), although the differences favoring r-hFSH remained even after stratification for several prognostic parameters, including age. In a post-hoc analysis of a prospective, randomized trial, Platteau et al. observed that HP-hMG was associated with significantly higher live birth rates than r-hFSH when used for COH in classic IVF cycles, although they reported similar efficacy in ICSI cycles [11]. However, when evaluated in a prospective, randomized IVF trial, there was no significant difference in ongoing pregnancy rates between HP-hMG and r-hFSH [12].

Although live birth rates were similar among the three treatment groups, the spontaneous abortion rate was significantly higher in the hMG group (Table 4). However, the significant difference in spontaneous abortion rates occurred only in the subgroup of women aged >35 years. One possible explanation was the finding that the mean age was significantly higher (1.0 years on average) in the hMG subgroup of women aged > 35 years compared with the other groups (Table 5).

In Brazil, the acquisition cost of r-hFSH is about 40-50% more expensive than the urinary menotropins; until now, this was an argument that supported the choice of urinary menotropins for COH. However, in our analysis, significantly lower doses of follitropin alfa per ICSI cycle were required to achieve similar clinical pregnancy and live birth rates. This observation was valid for the overall population as well as for the subgroups stratified for age, indication, and number of treatment attempts. Moreover, significantly higher doses of both menotropin products were necessary per pregnancy and per each live birth (20-45% higher) when compared with r-hFSH.

The explanations for these findings may relate to differences in the gonadotropin products due to improved processes available through utilization of recombinant technologies [13, 14]. Manufacturing improvements have resulted in significantly improved purity and specific activity, as well as batch-to-batch consistency of the follitropin alfa drug substance [14, 15]. Indeed, in clinical trials, increased efficiency was observed with follitropin alfa FbM in terms of a lower total dose and shorter treatment period when compared with FSH products [16‐19]. Also, as noted in the Methods, we routinely utilize a step-down ovarian stimulation protocol to avoid ovarian hyper-response. Because of the product improvements inherent with the follitropin alfa product, our clinicians are confident that dose adjustments with a decrement of 37.5 IU are clinically effective. The clinical effectiveness of 37.5 IU dosing has not been similarly demonstrated with either of the hMG products.

Similar live birth rates were achieved with hMG, HP-hMG and r-hFSH when used for COH after pituitary down-regulation with GnRH-a in ICSI cycles. Significantly lower total doses of r-hFSH were administered per cycle, compared with the menotropin products. Per each live birth, considerably higher doses of hMG and HP-hMG (45.3% and 19.8%, respectively) were used for ovarian stimulation. Thus, the greater purity and consistency of the r-hFSH FbM formulation resulted in higher treatment efficiency than urine-derived gonadotropins in terms of clinical pregnancy rates and live births.