Reproductive BioMedicine Online (2010) 21, 166 178

www.sciencedirect.com www.rbmonline.com

ARTICLE

HP-HMG versus rFSH in treatments combining fresh and frozen IVF cycles: success rates and economic evaluation
Jaro Wex-Wechowski Richard Kennedy c
a,*

, Ahmed M Abou-Setta a, Sandy Kildegaard Nielsen b,

a PharmArchitecture, Quatro House, Lyon Way, Camberley, Surrey GU16 7ER, UK; b Ferring International Centre, St-Prex, Switzerland; c Centre for Reproductive Medicine, University Hospitals Coventry and Warwickshire NHS, UK

* Corresponding author. E-mail address: jaro.wex@pharmarchitecture.com (J Wex-Wechowski). Jaro Wex-Wechowski is a physician with a degree in biology from Amherst College, USA. He gained his PhD in economics from Warsaw School of Economics via a study of nancing of health systems from the point of view of investment in cost-effective interventions. He has worked at health technology assessment and health insurance institutions and consultancies. Currently Jaro runs an independent health outcomes consultancy. In the area of reproduction, Jaro has worked on projects related to IVF, preterm birth and contraception. His major interests are evidence-based medicine, health systems nancing, pricing and reimbursement and health policy.

Abstract The economic implications of the choice of gonadotrophin inuence decision making but their cost-effectiveness in fro-

zen-embryo transfer cycles has not been adequately studied. An economic evaluation was performed comparing highly puried human menopausal gonadotrophin (HP-HMG) and recombinant FSH (rFSH) using individual patient data (n = 986) from two large randomized controlled trials using a long agonist IVF protocol. The simulation model incorporated live birth data and published UK costs of IVF-related medical resources. After treatment for up-to-three cycles (one fresh and up to two subsequent fresh or frozen cycles conditional on availability of cryopreserved embryos), the cumulative live birth rate was 53.7% (95% CI 49.358.1%) for HP-HMG and 44.6% (40.249.0%) for rFSH (OR 1.44, 95% CI 1.121.85; P < 0.005). The mean costs per IVF treatment for HP-HMG and rFSH were 5393 (53415449) and 6269 (62106324), respectively (number needed to treat to fund one additional treatment was seven; P < 0.001). With maternal and neonatal costs applied, the median cost per IVF baby delivered with HP-HMG was 11,157 (11,089 11,129) and 14,227 (14,18314,222) with rFSH (P < 0.001). The cost saving using HP-HMG remained after varying model parameters in a probabilistic sensitivity analysis. RBMOnline
2010, Reproductive Healthcare Ltd. Published by Elsevier Ltd. All rights reserved. KEYWORDS: cost-effectiveness, gonadotrophins, HP-HMG, IVF, live birth, rFSH

1472-6483/$ - see front matter 2010, Reproductive Healthcare Ltd. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.rbmo.2010.05.005

Cost-effectiveness of HP-HMG versus rFSH

167 evident that many couples are denied treatment for nancial reasons (Al-Inany et al., 2006; Collins, 2002; Jain et al., 2002; Staniec and Webb, 2007). The National Institute for Health and Clinical Excellence (NICE), the UK National Health Services (NHS) health guideline authority, has noted that according to the best available efcacy and cost-effectiveness evidence, up-to-three cycles should be provided to patients (NICE, 2004). The overall cost of these treatments to the NHS, and society as a whole, is expected to increase yearly in a linear fashion in parallel with the growth of the UK population. Even today. only approximately 25% of IVF cycles are being funded by the NHS (Kennedy et al., 2006) and only half of the couples with infertility seek treatment (Balen and Rutherford, 2007). Since the UK, like many other developed nations, has a moral obligation both to treat infertility and at the same time to address the rising costs related to this treatment, it is important to constantly re-evaluate the available options in order to determine the most appropriate and costeffective model for IVF treatment. The mere rationing of services and payments for infertility treatment is neither acceptable nor sensible. One alternative that has proven to be both practical and cost-effective is the use of excess, cryopreserved embryos from a fresh IVF cycle in order to increase the cumulative pregnancy and live birth rate per started cycle (Lukassen et al., 2005; Moustafa et al., 2008). Today, nearly every fertility clinic in the UK offers embryo cryopreservation to patients undergoing IVF. Excess embryos are cryopreserved to safely and efciently maximize the chance of conception through frozen-embryo transfers if fresh cycles fail (Lyerly et al., 2010). At the same time, cryopreservation prevents the need to transfer multiple embryos at one time. This option has become even more attractive in recent years in light of the trend to decrease the number of embryos transferred in an attempt to curb the high rate of multiple gestation pregnancies associated with assisted reproduction (Bissonnette et al., 2007). In a recent survey, 54% of respondents indicated that they were very likely to use cryopreserved embryos for future pregnancy attempts (Lyerly et al., 2010). In recent years, two large, prospective, randomized, multinational, non-inferiority trials were published demonstrating comparable clinical benet of highly puried human menopausal gonadotrophin (HP-HMG, Menopur; Ferring Pharmaceuticals, St-Prex, Switzerland) versus recombinant FSH (follitropin alpha, GONAL-F; Merck Serono, Geneva, Switzerland) (Andersen et al., 2006b; European and Israeli Study Group, 2002). Combined, the European and Israeli Study Group (EISG) and Menotrophin versus Recombinant FSH in vitro Fertilization Trial (MERiT) studies included almost 1500 women. Both were designed to compare the ongoing pregnancy rate in women undergoing down-regulation with a long gonadotrophin-releasing hormone (GnRH) agonist protocol and stimulated with either HP-HMG or rFSH. In both studies, the gonadotrophin dose was xed at 225 IU s.c. for the rst 5 days and then adjusted according to individual patient response. Several economic evaluations partially or completely built on the results of these two large trials have been published (Connolly et al., 2008; Lloyd et al., 2003; Wechowski et al., 2007, 2009), but only two were built on individual

Introduction
Assisted reproductive technologies, including IVF, have become a mainstream treatment option available to couples. Technological advances and a better understanding of human physiology and pathological processes that may lead to infertility have contributed to increased success rates (e.g., clinical pregnancies and live births) that are credited to IVF each year (Andersen et al., 2006a, 2007, 2008). At the same time, they have allowed clinicians to limit the incidence of complications due to fertility treatment by decreasing the number of embryos transferred and to provide early detection and management of complications (e.g., ovarian hyperstimulation syndrome, OHSS) (Land and Evers, 2003). The success of this form of medical treatment has become a routine practice with an almost guaranteed level of success, with 8090% of couples being successful in their rst year of treatment (Brosens et al., 2004). This has increased the attractiveness of IVF and has been reected in an exponential increase in fertility services usage with no contraction in sight (Lunenfeld et al., 2004). In the UK, approximately 15% of couples seek medical advice for infertility (Cahill and Wardle, 2002), with a lifetime prevalence of infertility ranging between 17.3% and 26.4% (Buckett and Bentick, 1997; Gunnell and Ewings, 1994). These projections are not expected to decline due to noted trends in lifestyle changes (e.g., increased age of marriage, sexually transmitted diseases, environmental pollution) in the general population (Balen and Rutherford, 2007). In fact, the prevalence of infertility increases to over 30% in the 3544-year-old group (Lunenfeld et al., 2004) and the lower fertility of older women is becoming increasingly signicant, as there is a trend over time for more women to postpone having their rst child until later in life (Commission of the European Communities, 2006). Success, however, comes at a price. The average cost of a successful IVF intervention leading to a live birth is variable; it is a cumulative gure built on multiple intricate decisions taken during the IVF procedure, including the type and quantity of medications (e.g., gonadotrophins) used for ovarian stimulation, oocyte retrieval, the fertilization procedure and laboratory-related activities, staff costs and costs related to cryopreservation (e.g., freezing, storing, thawing) in addition to post-transfer care. A signicant proportion of the share of the cost of IVF is attributed to medicines required for ovarian stimulation (Al-Inany et al., 2006). Therefore this clinical area is often targeted for effectiveness studies and cost-effectiveness analyses. The price of recombinant FSH (rFSH) is as high as three times that of urinary FSH (Al-Inany et al., 2006; Zwart van Rijkom et al., 2002). This has led to a vefold increase in total FSH expenditures in a 5-year period (i.e., 1995 to 2000), while the increase in usage of gonadotrophins was considerably smaller (Zwart van Rijkom et al., 2002). Due to the rising scal burden that infertility treatment has placed on national health programmes, standardization of the procedures using a cost-effectiveness-based approach has been promoted by many developed countries. In contrast, in countries where reimbursement of the cost of the IVF procedure is limited or not available through governmental programmes or insurance, it is unfortunately

168 patient data (Wechowski et al., 2007, 2009). The current study, supplementary to previous evaluations, sought to model the more clinically realistic scenario using available frozen-embryo transfers in addition to the fresh IVF cycles, in an attempt to determine the live birth rates per started cycle, as well as economic implication of choice of gonadotrophin. Naturally, these analyses take into consideration not only the costs of the respective medications in the UK, but also differences in live birth rates achieved from both the fresh and cryopreserved cycles, in addition to the expected availability of excess embryos for cryopreservationthawing for future transfers. The decision to use this pooled data in addition to the follow-up of cryopreservation transfers from the randomized women was made due to the robustness and pharmacological homogeneity of the preparations and dosages in the two trials. This allowed for a precise estimate of the cost difference between the two treatment options.

J Wex-Wechowski et al. tients, eligibility criteria, treatment protocols and primary endpoints were similar in both trials justifying the pooled analysis. Evidence on success rates following frozen cycles was obtained from a published report of follow-up of MERiT patients (Ziebe et al., 2007), where the live birth rate for both HP-HMG and rFSH was equal to 9%. For the scenario of one fresh and two frozen cycles success rates following frozen cycles were obtained both from MERiT and from the Human Fertilisation and Embryology Authority (HFEA) registry in the UK (Human Fertilisation and Embryology Authority, 2008). For the latter, success rates in frozen cycles were relative to fresh cycles. The thus obtained decrement was applied to both HP-HMG and rFSH treatments. When HFEA data were used, the advantage of HP-HMG established in clinical trials for fresh cycles was perpetuated into the subsequent frozen cycles. Due to their limited applicability in the analysis and potential bias, the HFEA data were not used in the economic evaluation, but were only considered part of sensitivity analysis for efcacy results. Drop-out rates following each cycle were taken from published sources (De Vries et al., 1999), as were the relative success rates for cycles within a treatment sequence (Meldrum et al., 1998).

Materials and methods

Overview
An evaluation comparing HP-HMG (Menopur) and rFSH (Gonal-F) in IVF treatments with cryopreserved embryos was conducted using a simulation model. The objective of the study was to determine: (i) efcacy expressed as number of patients with live births and number of babies born per patient initiating treatment; results of the simulation are an extension of the analyses presented in preceding papers (Andersen et al., 2006b; Platteau et al., 2004), which reported only outcomes from fresh IVF cycles; (ii) average IVF cost per treatment based on each of the gonadotrophin treatment options; the simulation results are reported as per patient treated, rather than per cycle, as it is the patients who were randomized in trials (Daya, 2003), and each treatment consists of a number of cycles contributing to the cumulative results per patient, therefore once a patient has a live birth, no further treatment (e.g., new fresh cycle or frozen-embryo transfer) is applied; (iii) cost per patient with a live birth and per baby for each treatment option; (iv) incremental cost-effectiveness expressed as the difference in cost of the two treatment options divided by the difference in number of live births or babies born per treated patient; and (v) impact of inclusion of maternal and neonatal costs arising from successful pregnancies, considering costs of multiple pregnancies.

Cost data
Resource use data was based on established UK treatment practices or items collected in the two randomized clinical trials. Unit costs per item were obtained from published UK sources. Treatment doses of HP-HMG and rFSH were sampled based on dosage reported in the two trials. Unit costs of 22.31 per vial of rFSH, containing 75 IU of FSH activity, and 13.65 per vial of HP-HMG, containing 75 IU of FSH activity and 75 IU of LH activity, were obtained from the British National Formulary (BNF; British Medical Association, 2008). In the UK centres, GnRH agonists were used according to the trial centre standard practice and therefore the costs of down-regulation were based on buserelin (Superfact; Sano-Aventis) 250 g/day for an average of 14 days. Also, to reect UK treatment practices, the costs of chorionic gonadotrophin alpha (Ovitrelle; Merck Serono) based on a single dose of 250 g at a cost of 33.31 were included in the cost calculations. Similarly, to reect UK treatment practices, the costs of vaginal progesterone gel (Crinone; Merck Serono) were included in the costs based on treatment ranging from 13 to 42 days after embryo transfer. Since daily application is indicated for 30 days after laboratory evidence of pregnancy, three packages with 15 dosages each per successful pregnancy were assumed at 32.73 per package. Based on expert advice all moderate and severe OHSS was assumed to lead to hospitalization at an average cost of 1032 per diagnosis. Frequency of OHSS was extracted from trial data. Per cycle frequency of hospitalization due to OHSS has previously been established at 1.9% for HP-HMG and 3.2% for rFSH based on an expert panel (Out et al., 1995) and used in previous economic evaluations (Sykes et al., 2001). In the Out et al. (1995) study, however, the HMG preparation used was not highly puried, nor is it commercially available at present. Therefore this study did not assume cross-efcacy or lack of it and it used

Clinical evidence
The modelled population was obtained by pooling patients from two prospective, randomized, multinational trials: EISG (IVF stratum, n = 255) and MERIT (all patients, n = 731) (Andersen et al., 2006b; European and Israeli Study Group, 2002; Platteau et al., 2008). Women, aged 1839, were recruited at 53 fertility clinics in 13 European countries and Israel. The entire modelled cohort was treated in fresh cycles with either HP-HMG (n = 491) or rFSH (n = 495) using a long GnRH agonist stimulation protocol. The primary causes of infertility in both studies were tubal and unexplained infertility and the baseline characteristics of pa-

Cost-effectiveness of HP-HMG versus rFSH HP-HMG clinical trial data in the analyses instead. There was no difference in OHSS between the two treatment options, hence the actual frequencies were used: 1.83% and 1.81% for HP-HMG and rFSH, respectively. One of the authors (RK) conrmed from his clinical experience that the previously reported assumptions were applicable to this study, although to account for the fact that patients with moderate OHSS are hospitalized for shorter periods, it was assumed that average hospitalization did not exceed 5 days, a non-elective trim point (HRG Tariff 2007). Infrequent occurrence of OHSS, however, had virtually no impact on the results of the analyses. Cost of patient visits was assumed to be included in costs of retrieval, fertilization and embryo transfer. Specialist consultation costs were added during stimulation and for pregnancy determination. Cost of delivery and neonatal care were considered separately for singletons, twins and triplets (Ledger et al., 2006). Cost and outcome consequences of failed treatments and the probability of adoption were not considered. The perspective of the study was that of the NHS payer and non-medical costs and costs of time and loss of productivity were not incorporated in the analysis. Since fewer cycles would be required for HP-HMG than rFSH to achieve one delivery and one birth (baby), the exclusion of social costs had a conservative effect likely to underestimate the cost advantage of HP-HMG. Costs were indexed to 2008, with cost of drugs remaining unchanged as of 1 May 2009, and no discounting was applied due to the short time horizon employed. All input variables for costs used in the model are summarized in Table 1.

169 events and therefore were not considered explicitly in the model. The logical ow of the model is illustrated in Figure 1; the process is briey described below. Each patient entered the simulated treatment pathway in the stimulation state. Treatment with gonadotrophins was administered in this state, accounting for prior pituitary down-regulation, as described previously (Andersen et al., 2006b; European and Israeli Study Group, 2002). Patients underwent ovulation induction and entered the oocyte retrieval state. Numbers of oocytes retrieved and fertilization rates were retrieved from the trials to allow for an accurate modelling of the effects of each gonadotrophin preparation. Following fertilization, for those patients with viable embryos available, embryo transfer was performed which led to the stage of pregnancy which resulted in either pregnancy loss or live birth, single or multiple delivery. In case of unsuccessful outcome, the patient could discontinue treatment (drop-out) or start another cycle. The second cycle could be either fresh or frozen, depending on the analysed scenario. In the base case scenario, patients were offered second fresh cycles, unless cryopreserved embryos were available. The availability of cryopreserved embryos was modelled as directly related to the number of oocytes retrieved. It was assumed that at least three oocytes had to be retrieved for a cryopreserved embryo to be available, assuming conservatively, for both treatment options, 33.9% minimum quality embryos per oocyte retrieved (Ziebe et al., 2007). When the relative share of single- and double-embryo transfers, as well as fertilization rates were used from MERiT (Andersen et al., 2006b), the required minimum number of oocytes retrieved was also three. The number of required oocytes was tested in a sensitivity analysis in the range from three to seven, producing a trend for favouring HP-HMG with increasing requirement for the number of oocytes retrieved. This trend resulted from greater impact of drug cost with fewer frozen cycles. Acknowledging that possible variation in clinical practice with regard to singleversus multiple-embryo transfer, as well as efciency of fertilization and cryopreservation, might necessitate considerably more oocytes to be retrieved per frozen cycle, a conservative approach favouring rFSH was selected. The number of embryos transferred was assumed to be the same for both groups and was not modelled to be associated with success rates, which were determined as described above. The transfer of cryopreserved embryos in the second cycle diminished their pool available for the third frozen cycle. The third cycle was offered following previous unsuccessful attempts, subject to drop out; frozen embryos were transferred subject to availability.

The simulation model

A discrete event simulation was designed to model clinical outcomes and-related costs and coded using Visual Basic for Applications in Microsoft Excel. Probabilities based on individual patient data from the combined trial cohort were simulated for each clinical event, using computer-generated random numbers representing chance events (Higgins and Green, 2008). In such simulations, also known as Monte Carlo simulations, results from numerous runs are averaged over all patients, approximating the results that would be obtained if average probabilities, identical for each patient were used. In simulations, the events do not need to be synchronized, as in Markov models, allowing greater exibility and more accurate representation of the patient pathway and thus reecting clinical reality with greater accuracy. For the analyses, a total of 30,000 patients were simulated, a cohort sufcient for non-signicant differences during multiple model runs. The following treatment events (stages) in the rst (always fresh) treatment cycle were modelled from the combined clinical cohort: stimulation, oocyte retrieval, embryo transfer, pregnancy and live birth. Modelled treatment was consistent with the procedures performed in the clinical trials and the patient transitions between stages of the model were determined by pooled patient-level data from the two clinical trials. The number of oocytes retrieved and number of twins and triplets were also obtained from the trials data. Probabilities of fertilization and implantation success were accounted for by subsequent

Statistical analysis
Efcacy results were simulated using available case intention-to-treat data (Higgins and Green, 2008). To test robustness of the results, the last observation carried forward (LOCF) method for imputing missing values was also used; values for 57 patients were imputed in the HP-HMG and for 55 in the rFSH arm. Only the available case results are reported, as the LOCF-based results were not signicantly different and did not affect the conclusions. Chi-squared test of association was used to test signicance of efcacy

170
Table 1 Variable Cost of Gonal-F (rFSH) per IU Cost of Menopur (HP-HMG) per IU Median dose rFSH Cost variables used in the economic evaluation of simulation. Value 0.298 0.182 Sample trial data Uncertainty interval 0.2380.357 0.1460.218 12756525 IU Source British Medical Association (2008) British Medical Association (2008) Lloyd et al. (2003) and Andersen et al. (2006b) Lloyd et al. (2003) and Andersen et al. (2006b) Andersen et al. (2006b) and British Medical Association (2008) Andersen et al. (2006b) and British Medical Association (2008) British Medical Association (2008) Comments

J Wex-Wechowski et al.

Based on price of 22.31 per vial containing 75 IU of rFSH Based on price of 13.65 per vial containing 75 IU of FSH and 75 IU of LH Dose was bootstrapped from pooled MERiT and EISG trial data due to right skewness of the distribution (median 2250 IU)

Median dose HPHMG Cost of Buserelin

Sample trial data 7.54

12756825 IU

6.039.05

Superfact (Sano-Aventis), unit cost of 11.85 per 5.5 mg vial. Dosing based on local practice, 250 lg/day for 14 days Chorionic gonadotrophin-alpha (Ovitrelle; Merck Serono) was costed as a single dose at 250 lg Vaginal progesterone gel (Crinone; Merck Serono) was costed after embryo transfer during 1342 days. Since daily application is indicated for 30 days after laboratory evidence of pregnancy, three packages with 15 dosages each per successful pregnancy were assumed at 32.73 per package UHCW cost inated to 2008 price UHCW cost inated to 2008 price UHCW cost inated to 2008 price Inated to 2008 price. All moderate and severe OHSS was assumed to lead to hospitalization Inated to 2008 price. Based on HCG test cost of 14.83 and specialist consultation at 34 Inated to 2008 price. Based on ultrasound cost of 38 and specialist consultation at 34 Inated to 2008 price. Costs of early and late pregnancy loss were not differentiated

Cost of chorionic gonadotrophinalpha Luteal support

33.31

26.6539.97

24.06

19.2528.87

Cost of retrieval Cost of embryo transfer Cost of fertilization Hospitalization cost for OHSS Cost of clinical pregnancy determination Cost of ongoing pregnancy determination Cost of pregnancy loss

795 520 246 1079

636954 416624 197295 8631295

Lloyd et al. (2003) Lloyd et al. (2003) Lloyd et al. (2003) Daya et al. (2001)

49

3959

72

6579

281

225337

Daya et al. (2001) and Curtis et al. (2005) Daya et al. (2001) and Curtis and Netten (2005) NICE (2004)

EISG = European and Israeli Study Group; HCG = human chorionic gonadotrophin; HP-HMG = highly puried human menopausal gonadotrophin; MERiT = Menotrophin versus Recombinant FSH in vitro Fertilization Trial; OHSS = ovarian hyperstimulation syndrome; rFSH = recombinant FSH; UHCW = University Hospitals of Coventry and Warwickshire.

outcomes. Odds ratios and their condence limits were calculated using the asymptotic method (Bland and Altman, 2000). Independence of babies from multiple pregnancies was assumed, which has no effect on calculated odds ratio (Gates and Brocklehurst, 2004). Simulation output cost variables were not normally distributed; hence the Wilcoxon ranks sum test was used to test the signicance of mean and median results. In addition, the results were veried using a permutation test on

means and medians of samples obtained by bootstrapping the cost results from the model. Also condence intervals were produced by bootstrapping. Bootstrapping estimates the sampling distribution of a statistic through a large number of simulations, employing sampling with replacement from the original data (Briggs et al., 1997). Samples of sizes equal to the trial groups were drawn from the simulated parent population of 50,000 patients. Mean values were used for summary reporting,

Cost-effectiveness of HP-HMG versus rFSH

171 butions were used to estimate condence intervals for medians (Conover, 1980). Considerable skewness of the gonadotrophin dose distribution was addressed by sampling the original trial data. Comparative cost-effectiveness of gonadotrophins was expressed using an incremental costeffectiveness ratio (ICER). The ICER is a comparative metric used to represent the additional cost required to achieve an additional unit of clinical effectiveness between two different treatment options (Gold et al., 1996). Probabilistic sensitivity analysis was performed to address uncertainty in parameters used in the model. Based on accepted practices, values for parameters were sampled 1000 times from 95% CI based on probabilities of success at each stage of simulation and from uncertainty intervals for costs, judged to conservatively approximate the 95% CI. Results from the probabilistic sensitivity analysis were summarized using cost-effectiveness acceptability curves (Fenwick et al., 2004), which present the probability that a given intervention is more cost-effective than the alternative treatment as a function of ceiling ratio or willingness-to-pay for one unit of effectiveness.

Results
Figure 1 Flow diagram showing treatment pathway with simulation events. avail. = available.

Efcacy
Based on the combined data from the MERiT and EISG trials, the cumulative live birth rate resulting from treatments consisting of a sequence of one fresh and up to two either fresh or frozen cycles, conditional on availability of frozen embryos, was 53.7% (95% CI 49.358.1%) for HP-HMG and 44.6% (95% CI 40.249.0%) for rFSH (OR 1.44, 95% CI 1.12 to 1.85; P = 0.004; Table 2). When the minimum number

except for costs inclusive of maternal and neonatal costs, where means were inappropriate due to skewed irregular distribution attributed to costs associated with twins and triplets and where medians were used instead. Two-sided 95% CI were obtained based on cost distributions from bootstrap as the 2.5% and the 97.5% percentiles; binomial distri-

Table 2

Success rates in different simulation scenarios. HP-HMG rFSH P-value Odds ratio

Scenarios and outcomes One fresh + up to two frozen cycles, conditional on availability of oocytes for cryopreservation (minimum three oocytes) Live birth rate (%) Babies per 1000 treatments One fresh + one frozen cycle Live birth rate (%) Babies per 1000 treatments One fresh + two frozen cycles (frozen success rate from MERiT follow-up) Live birth rate (%) Babies per 1000 treatments One fresh + two frozen cycles (relative frozen success rate from HFEA) Live birth rate (%) Babies per 1000 treatments

53.7 (49.358.1) 679 (639720) 32.6 (28.436.7) 414 (370458)

44.6 (40.249.0) 580 (536623) 27.0 (23.130.1) 350 (308392)

0.004 0.001 NS 0.038

1.44 (1.121.85) 1.53 (1.181.99) 1.31 (0.991.72) 1.31 (1.011.70)

36.4 (32.140.6) 460 (416504)

32.1 (28.036.2) 419 (376463)

NS NS

1.21 (0.931.57) 1.18 (0.921.52)

49.2 (44.753.6) 626 (584669)

39.9 (35.644.2) 520 (476564)

0.003 <0.001

1.46 (1.131.87) 1.55 (1.201.99)

Values in parentheses are 95% CI. EISG = European and Israeli Study Group; HFEA = Human Fertilisation and Embryology Authority; HP-HMG = highly puried human menopausal gonadotrophin; MERiT = Menotrophin versus Recombinant FSH in vitro Fertilization Trial; NS = not statistically signicant; rFSH = recombinant FSH.

172 of oocytes required for one frozen cycle was changed to seven, the respective live birth rates were 54.7% and 45.2% and the difference remained statistically signicant (P < 0.005). Using live birth rates from individual frozen cycles, in the HFEA registry data, also demonstrated an advantage of stimulation with HP-HMG (54.2% versus 44.8%, respectively, P < 0.005). When the success rate was measured by the number of babies born, accounting for multiple births, the difference in rates was also statistically signicant and also in favour of HP-HMG (P < 0.001; Table 2). The required number of patients needed to be treated to achieve one additional live birth when using HP-HMG instead of rFSH varied in different scenarios from 11 to 24; the number of babies needed to be treated varied from 10 to 25. In the more theoretical scenario, assuming unconditional availability of frozen embryos and involving one fresh and exactly one frozen cycle, the cumulative live birth rate was 32.6% for HP-HMG and 27.0% for rFSH and the number of babies born was 414 and 350 per 1000 patients treated, respectively (P = 0.038; Table 2). When exactly two frozen cycles following a fresh cycle were allowed, the advantage of using HP-HMG remained, although statistical signicance of the difference in success rates depended on assumptions used in the model. When the success rate for frozen cycles from the MERiT trial was used, the difference was not signicant, while using the HFEA registry data resulted in a high level of signicance for both live births and babies (P = 0.003 and P < 0.001, respectively; Table 2). Simulations based on evidence from MERiT only reected the trend for advantage of using HP-HMG versus rFSH established in the trial. The cumulative live birth rate varied in different scenarios: 32.454.0% for HP-HMG and 29.3 47.3% for rFSH.

J Wex-Wechowski et al.

Costs
Based on the combined data from the MERiT and EISG trials, total mean IVF costs per patient starting the treatment after one fresh and up to two fresh or frozen treatment cycles, conditional on availability of frozen embryos, for HP-HMG and rFSH were 5393 (95% CI 53415449) and 6269 (95% CI 62106324), respectively (P < 0.001; Table 3). The mean per treatment cost savings resulting from using HP-HMG instead of rFSH was 876. This saving would allow nancing of an additional treatment for every seven HP-HMG treatments delivered. The total cost of gonadotrophin was 39.5% lower for HP-HMG than for rFSH, i.e., 990 (95% CI 9631018) versus 1637 (95% CI 15981679; P < 0.001; Table 3). Other costs related to IVF procedures, inclusive of OHSS treatment costs, were comparable (3283 versus 3489, respectively), as were other costs related to stimulation (831 versus 843, respectively). When maternal and neonatal costs were attributed to live birth rate data, accounting for occurrence and costs of multiple pregnancies, the median costs per patient starting the treatment with HP-HMG and rFSH were 7571 and 8278, respectively (P < 0.01). Results for scenarios involving one fresh plus one or two frozen cycles are presented (Table 4). Across the scenarios, the cost saving resulting from using HP-HMG instead of rFSH would allow funding of an additional treatment for every 6 11 treatments delivered. Costs per patient starting the treatment were consistently lower in treatments involving stimulation with HP-HMG. The results for simulation using MERiT data only showed a similar pattern.

Table 3 Treatment costs for HP-HMG and rFSH treatments using one fresh and up to two frozen cycles, based on availability of oocytes. Cost outcome Cost () HP-HMG Gonadotrophin costs per patient treatment Mean Median Total costs per patient treatment Mean Median Cost per live birth excluding maternal and neonatal costs Mean Median Cost per baby including maternal and neonatal costs Mean Median rFSH

990 (9631018)a 988 (969983) 5393 (53415449)a 5369 (51445158) 10,046 (923811,002)a 9648 (95839608)b 12,635 (987415,156)a 11,157 (11,08911,129)b

1637 (15981679)a 1638 (16511673) 6269 (62106324)a 6244 (59065973) 14,055 (12,73915,654)a 13,341 (13,24313,393)b 15,563 (13,27919,875)a 14,227 (14,18314,222)b

Values in parentheses are 95% CI. HP-HMG = highly puried human menopausal gonadotrophin; rFSH = recombinant FSH. a Signicant differences between HP-HMG and rFSH (P < 0.001; Wilcoxon rank sum test), despite instances of overlap of condence intervals. In permutation tests, condence intervals for the difference in means and medians did not contain 0, indicating signicance in all comparisons. b Signicant differences between HP-HMG and rFSH (P < 0.001; permutation test). Condence intervals for medians provided for completeness of statistical measures; they have no clinical interpretation when the median value lies outside the interval; condence intervals for differences obtained in permutation tests did not include zero, indicating signicance.

Cost-effectiveness of HP-HMG versus rFSH

Table 4 Treatment costs for HP-HMG and rFSH treatments using 1 fresh and 1 or 2 frozen cycles. Scenarios and cost outcomes Cost () HP-HMG One fresh + one frozen cycle Gonadotrophin costs per patient treatment Mean total costs per patient treatment Mean cost per live birth excluding maternal and neonatal costs One fresh + two frozen cycles Gonadotrophin costs per patient treatment Mean total costs per patient treatment Mean cost per live birth excluding maternal and neonatal costs rFSH

173

472 3562 10,933 471 4242 11,659

734 3896 14,426 737 4622 14,388

All differences between HP-HMG and rFSH are signicant (P < 0.001; Wilcoxon rank sum test). HP-HMG = highly puried human menopausal gonadotrophin; rFSH = recombinant FSH.

Cost-effectiveness
Based on the combined data from the MERiT and EISG trials, total mean IVF costs per live birth after one fresh and up to two frozen treatment cycles, conditional on availability of frozen embryos, for HP-HMG and rFSH were 10,046 (95% CI 923811,002) and 14,055 (12,73915,654), respectively (P < 0.001; Table 3). When maternal and neonatal costs were added, the respective median costs per live birth were 12,635 with HP-HMG and 15,563 with rFSH. Table 3 shows the average cost-effectiveness results expressed as mean and median cost. The ICER was negative for both IVF cost per live birth and IVF cost per baby, indicating that HP-HMG is less costly and more effective (dominant) versus rFSH (Table 5). The cost saving combined with superior effectiveness remained when maternal and neonatal costs were added to the cost of IVF. This pattern was observed in other analysed scenarios.

assumed to be zero, the cost per live birth when treating with rFSH would still be higher than that of treating with HP-HMG, because of other costs (e.g., procedures, tests, etc.) incurred in the treatment needed to produce one live birth with rFSH (Figure 3). An equivalent total cost per IVF treatment would be achieved at the cost of rFSH 1.86 per vial (data not shown). This corresponds to a price of rFSH equal to 12% of the published BNF price for follitropin alpha.

Probabilistic sensitivity analysis

The probabilistic sensitivity analysis supported the conclusion drawn from the base case that HP-HMG was cost saving compared with rFSH. The mean and median value of ICER was negative both for cost per delivery and cost per baby. The cost-effectiveness acceptability curve showed that HP-HMG was cost saving in 100% of patients, both exclusive and inclusive of neonatal and maternal costs, when willingness-to-pay for live birth or baby (ceiling ratio) was varied from 0 to 20,000.

Changes in drug acquisition costs

The impact of varying the acquisition costs of rFSH on comparative costs per live birth after one treatment involving one fresh and up to two frozen cycles was tested in the model. The results indicate that an equivalent cost per live birth for both gonadotrophins cannot be achieved, regardless of the cost of rFSH. To achieve an equivalent cost per live birth, the price of rFSH would have to be negative (Figure 2). Even if the drug acquisition cost of rFSH was

Health system implications

The analysis also estimated the likely budget impact of switching patients currently treated with rFSH to HP-HMG. Based on the total number of 34,855 women treated annually with IVF (HFEA 2009) and assuming 50% patients on rFSH, the total savings would be 15,266,490. This is a conservative estimate assuming all women received only one

Table 5 Incremental ratios (highly puried human menopausal gonadotrophin versus recombinant FSH) after treatment with one fresh and up to two frozen cycles, conditional on availability of oocytes. Cost outcome Incremental Incremental Incremental Incremental cost cost cost cost per per per per live birth baby delivery including maternity and neonatal costs baby including maternity and neonatal costs Incremental value () 9632 8856 4938 4540

Negative values indicate a cost saving with highly puried human menopausal gonadotrophin.

174

J Wex-Wechowski et al. to become the mainstream primary outcome of trials of effectiveness in assisted reproductive technologies. This new-found revelation is supported by the observations that previously accepted benchmarks (e.g., biochemical and clinical pregnancy rates) did not properly relate to the patients needs. A new trend that is also evolving is the cost burden to individuals, organizations, governments and society as a whole as a result of infertility treatment. With budget constraints and ination in medical costs, more weight is being placed on incorporating cost and effectiveness in any decision regarding infertility management (Mladovsky and Sorenson, 2009). This also includes calls for the need to provide cost-effective treatment in medical care generally, as well as well as infertility practice. IVF has become medically and publicly accepted as a successful option for treatment of infertile couples and patients seeking protection from the transmission of sexually transmitted diseases (e.g., HIV) and genetic disorders. Over the past decade, breakthroughs in biomedical engineering and urinary purication processes have produced purer, more consistent products for ovarian stimulation. This has led to improved clinical pregnancy rates and live birth rates that were not previously attainable. At the same time, the improved pregnancy rates have carried the burden of a higher than normal chance of a multiple gestation and rising cost of infertility treatment. This has resulted in political calls for the use of milder ovarian stimulation protocols, the transfer of fewer embryos and the use of more cost-effective medications (Van Voorhis, 2006). Historically, gonadotrophins used for ovarian stimulation were processed from the urine of menopausal women (e.g., human menopausal gonadotrophins; HMG). In the mid1990s, rFSH was marketed but it carried a signicantly higher price per IU of FSH. To offset this price difference, investigators demonstrated that due to the higher pregnancy rates and less need for gonadotrophins for stimulation, the recombinant option was more cost-effective (Daya et al., 2001). Recently, traditional HMG has been replaced by HP-HMG, which contains a higher concentration of human chorionic gonadotrophin (HCG) (Wolfenson et al., 2005). Although all HMG preparations contain HCG, HP-HMG contains more HCG than traditional HMG, and lower lLH concentrations. This increased HCG component in HP-HMG provides most of the LH activity required for oocyte development (Wolfenson et al., 2005). Furthermore, this new balance has been shown to induce a different follicular development prole than traditional HMG products (Filicori et al., 2005; Platteau et al., 2006) and rFSH (Smitz et al., 2007). Today, HP-HMG and rFSH are two of the commonest medications prescribed for ovarian stimulation. Both products are proven in clinical trials to be both efcient and produce similar pregnancy and live birth rates, although pooled analyses revealed signicantly higher success rates with HP-HMG (Arce and Srensen, 2005; Wechowski et al., 2009). A recent meta-analysis found a trend for superior efcacy of HP-HMG in IVF, concluding that the former should be preferred on grounds of efcacy (Al-Inany et al., 2009). Even so, the uptake of a more cost-effective strategy will allow individuals more access to limited funds by increasing the number of cycles that an organization can

Figure 2 Cost per live birth as a function of acquisition cost of recombinant FSH (rFSH; solid line) and highly puried human menopausal gonadotrophin (HP-HMG; dashed line).

Figure 3 Cost per live birth disaggregated into cost of gonadotrophin and other IVF costs.

treatment in a given year. With the savings, an additional 2831 patients could be treated with IVF, leading to 1922 more babies being born. As HP-HMG also leads to higher success rates, regardless of the cost of treatment, a further 1725 babies would be born if HP-HMG is chosen, giving a combined gure of 3647 additional babies. If the gures are scaled down to actual number of IVF babies born annually, the total system savings would be 8125,255 with 1941 additional babies born. The adjusted gures, although more realistic, also reect limited access to IVF as many patients would not be offered the recommended up-to-three cycles, which were modelled in this study.

Discussion
The results of this simulation model using a sequence of one fresh and up to two fresh or frozen cycles, conditional on the availability of cryopreserved embryos, demonstrated not only a superior cumulative live birth rate for HP-HMG compared with rFSH, but also showed that the mean costs per IVF treatment were signicantly less for HP-HMG. When maternal and neonatal costs were applied, the median cost per IVF baby delivered was still signicantly less with HP-HMG. This cost saving from using HP-HMG depicted in this model would allow an additional treatment cycle for every seven patients treated. The exact denition of success in assisted reproduction has been debated for years. Today, live birth has evolved

Cost-effectiveness of HP-HMG versus rFSH afford to provide. With the non-inferior results of both these gonadotrophin preparations, the burden of deciphering this equation lies with cost-effectiveness modelling. Lloyd et al. (2003) performed a cost-minimalization analysis to determine the cost of achieving pregnancy with HP-HMG versus rFSH. Using data from EISG (2002), they concluded that HP-HMG was less expensive per cycle, translating into a 13% increase in the number of cycles that could be offered. Wechowski et al. (2007) performed an economic evaluation using a discrete event simulation model to assess treatment costs in relation to live births for both treatments based on published UK costs. After one cycle, the mean costs per IVF treatment for HP-HMG and rFSH were 2396 (95% CI 23832414) and 2633 (26152652), respectively. The average cost saving of 237 per IVF cycle using HP-HMG allows one additional cycle to be delivered for every 10 cycles. Based on the results of the recently published meta-analysis by Coomarasamy et al. (2008) comparing HMG and rFSH, Connolly et al. (2008) constructed a simulated decision tree model. Gonadotrophin costs were based on Menopur and Gonal-F prices in Belgium. After one fresh and one cryopreserved cycle, the average treatment cost with HP-HMG was lower than with rFSH (3635 and 4103, respectively). This was the rst model to take into consideration the cumulative pregnancy rate following the use of cryopreserved embryos produced from the original fresh cycle. Even so, it should be noted that the aforementioned meta-analysis included both traditional HMG and HP-HMG and both recombinant alpha- and beta-FSH preparations. These clinical heterogeneities would not be expected to completely and accurately model the effect of HP-HMG compared with recombinant alpha-FSH. Specically, in light of the higher trend of pregnancy rates with HP-HMG compared with traditional HMG (Al-Inany et al., 2009), they could have lead to an underestimation of the effectiveness of HP-HMG. An economic evaluation was recently published that used a simulation model in order to determine the cost per IVF cycle and cost per live birth for HP-HMG and rFSH, based a pooled analysis integrating the individual follow-up data obtained from two large randomized trials that compared identical preparations and followed a similar stimulation protocol (MERiT and EISG) (Wechowski et al., 2009). Using price data from the UK, HP-HMG was demonstrated to be more effective and less costly for both IVF cost per live birth and for IVF cost per baby. The results of the current cost-effectiveness analysis demonstrate that HP-HMG is cost saving compared with recombinant alpha-FSH in terms of live births and babies born after both fresh and combined fresh- and frozenembryo transfers, when available. This data is in line with previous work (Wechowski et al., 2007, 2009) and that of other investigators (Connolly et al., 2008). The importance of adding frozen-embryo cycles to the current analysis provides a more robust and realistic approach to determining the cost of treatment per cycle as opposed to the use of fresh cycles only. This is in line with the ethical and political determination of curbing the iatrogenic complication of multiple gestations produced from the transfer of high numbers of embryos by using a policy of less embryos per transfer coupled with more cycles of

175 embryo transfers (American Society for Reproductive Medicine, 2004; European Society of Human Reproduction and Embryology, 2000; Thurin et al., 2004). The policy of multiple single-embryo transfers (Templeton and Morris, 1998; Veleva et al., 2006) or the use of one fresh plus one frozen-embryo transfer (Moustafa et al., 2008) has been demonstrated to be as effective as one double-embryo transfer. This has led to a marked decrease in the number of embryos per transfer and, in turn. a greater number of embryos being cryopreserved for future use. One important issue to consider when choosing the type of gonadotrophin is the quality of embryos produced following stimulation. rFSH has been demonstrated to produce higher numbers of oocytes per stimulated cycle, but at the same time the embryo quality per oocyte retrieved has been proven to be higher with HP-HMG (Ziebe et al., 2007). This is of great importance, since embryo quality prior to cryopreservation has been demonstrated to affect the live birth rate after a freezethaw transfer cycle (Salumets et al., 2006). In addition, cryopreserved embryos originating from conception cycles achieve double the implantation and pregnancy rates of those obtained from unsuccessful cycles (El-Toukhy et al., 2003). These points add further advantage to the use of HP-HMG. Currently, there is limited data from randomized trials on the outcome of cryopreserved embryos following ovarian stimulation using HP-HMG compared with rFSH. A meticulous search of the literature (Medline, Embase and Central) revealed data on the developmental outcome of frozen embryos and frozen-embryo transfer cycles from only the MERiT (Ziebe et al., 2006, 2007) and First IVF-ICSI Cycle Recombinant FSH vs. Menotropin (FIRM) studies (Hompes et al., 2008). Data from the MERiT study was only available for a 1-year follow-up period in which 178 patients (24.35%) of the original cohort underwent thawing of cryopreserved embryos. The results demonstrated more preferable outcomes for embryos thawed in the HP-HMG group compared with the rFSH group, with a higher frequency of embryos with >50% intact blastomeres and resumed mitosis prior to transfer. However, the pregnancy rate following the rst frozen cycle was the same for both groups (9%) and no information on the cumulative pregnancy rate was presented. The FIRM trial provided similar clinical results to those in the MERiT study, with a major difference in that no data was available on the quality of the embryos post thawing and that information on the cumulative pregnancy rate was presented. The ongoing pregnancy rate following frozen-embryo transfers was non-signicantly higher in the HP-HMG group (5/20, 25%) than in the rFSH group (4/35, 11.43%). This increased the 1-year cumulative ongoing pregnancy rate from 26.3% to 28.0% in the HP-HMG group and 25.2% to 26.6% in the rFSH group. The current studys model conservatively assumed equal success rates following frozen cycles, based on the MERiT follow-up data (Ziebe et al., 2007). Use of the relative fresh/frozen success rate obtained from the HFEA registry data may incorporate the above points related to embryo quality, as it preserves the advantage of HP-HMG, as shown in the MERiT and EISG trials by applying to both drugs the same percentage decrease in frozen versus fresh cycles. Although both the approaches based on Ziebe et al. (2007) and HFEA demonstrate superior success rates on HP-HMG,

176 only incorporating the HFEA data produce statistical significance. The current study took a conservative approach, using only clinical trial data in the economic evaluation. The majority of assisted reproductive technologies treatments in the UK are IVF, rather than intracytoplasmic sperm injection (ICSI), with even fewer ICSI treatments funded by the NHS. This study evaluated efcacy and cost-effectiveness of IVF, as clinical trial data on live births following ICSI were unavailable. Using combined IVF and ICSI data would not be an accurate representation of either procedure, since only IVF procedures were studied in MERiT. While comparable efcacy of HP-HMG and rFSH has been demonstrated in a recent meta-analysis (Al-Inany et al., 2009), the economic consequences of using either alternative would depend on unit costs of gonadotrophins. A costminimization analysis could be expected to show superior value for money of a less expensive drug, which is HP-HMG. The analyses were based on evidence from clinical trials of patients whose age ranged from 18 to 39 years. While treatment effectiveness is lower in older patients, there is no evidence suggesting that the difference between gonadotrophins would be different in different age groups. The preliminary analysis of patient-level data from the combined MERiT and EISG trials showed that when younger patients were excluded (25 years old and younger), the advantage of HP-HMG was even more pronounced than on average, with 13% difference in success rates. It can also be expected that in older patients with decreasing embryo quality, more fresh cycles would have to be performed to achieve comparable results, with would further favour HP-HMG as the less costly option. One of the limitations of this study stems from unavailability of actual market prices of gonadotrophins, which were approximated using BNF prices. BNF price is considered an established standard in economic evaluations in the UK when more accurate estimates are not available. Indeed, for illustrative purposes, the costs must be based on an objective source rather than a more subjective assessment of possible costs based on volume of purchases with short- and long-term sales targets. Incorporating discounts from non-randomly selected clinics, even if obtainable, could introduce bias, as discounts can be local and temporary, reecting marketing strategies, and distort the true average cost of drugs. The BNF price, even if not the most accurate, offers transparency and equivalency across centres. Any provider able to purchase drugs at a discount would be able to factor in price information while interpreting these results, though sensitivity analysis clearly demonstrated that cost of gonadotrophins should have no bearing on purchasing decision, use of HP-HMG being cost saving regardless of the cost of rFSH. In addition, to account for price variation, unit costs of gonadotrophins were varied within a 20% uncertainty interval in the probabilistic sensitivity analysis. Furthermore, from the perspective of private patients, rather than private providers, costs of gonadotrophins appear to be in line with BNF, if not greater, which is a conservative assumption in favour of HP-HMG. This evaluation was conducted from the perspective of the NHS payer, although it is believed that results are broadly applicable to the private sector payers as well. Higher success rates with HP-HMG require fewer cycles, which would lead to even greater savings than in the NHS,

J Wex-Wechowski et al. due to greater per procedure costs in private clinics. Higher success rate would also translate into competitive advantage of clinics using HP-HMG. In conclusion, the results of this economic analysis demonstrate the superior cost-effectiveness of HP-HMG to produce live births over rFSH in women undergoing conventional IVF and, when available, frozen-embryo transfer cycles. Since this work more realistically models the actual situation in patients undergoing infertility management, it is believed that it will have a major implication on the decision-making process when choosing the appropriate gonadotrophin for ovarian stimulation.

Acknowledgements
JW and AA were provided an unrestricted research grant from Ferring Pharmaceuticals in support of this project.

References
Al-Inany, H.G., Abou-Setta, A.M., Aboulghar, M.A., Mansour, R.T., Serour, G.I., 2006. HMG versus rFSH for ovulation induction in developing countries: a cost-effectiveness analysis based on the results of a recent meta-analysis. Reprod. Biomed. Online 12, 163169. Al-Inany, H.G., Abou-Setta, A.M., Aboulghar, M.A., Mansour, R.T., Serour, G.I., 2009. Highly puried hMG achieves better pregnancy rates in IVF cycles but not ICSI cycles compared to recombinant FSH: a meta-analysis. Gynecol. Endocrin. 27, 17 (Epub ahead of print). American Society for Reproductive Medicine, 2004. Guidelines on number of embryos transferred. Fertil. Steril. 82, 773774. Andersen, A.N., Gianaroli, L., Felberbaum, R., de Mouzon, J., Nygren, K.G., 2006a. The European IVF-monitoring programme (EIM) for the European Society of Human Reproduction and Embryology (ESHRE), 2006a. Assisted reproductive technology in Europe, 2002. Results generated from European registers by ESHRE. Hum. Reprod. 21, 16801697. Andersen, A.N., Devroey, P., Arce, J.C., 2006b. Clinical outcome following stimulation with highly puried hMG or recombinant FSH in patients undergoing IVF: a randomized assessor-blind controlled trial. Hum. Reprod. 21, 32173227. Andersen, A.N., Goossens, V., Gianaroli, L., Felberbaum, R., de Mouzon, J., Nygren, K.G., 2007. Assisted reproductive technology in Europe, 2003. Results generated from European registers by ESHRE. Hum. Reprod. 22, 15131525. Andersen, A.N., Goossens, V., Ferraretti, A.P., et al., 2008. European IVF-monitoring (EIM) Consortium; European Society of Human Reproduction and Embryology (ESHRE), 2008. Assisted reproductive technology in Europe, 2004: results generated from European registers by ESHRE. Hum. Reprod. 23, 756771. Arce, J., Srensen, P., 2005. A pooled analysis of two large randomized trials comparing ongoing pregnancy rates with highly puried menotropin and recombinant FSH in IVF cycles. Fertil. Steril. 84 (Suppl. 1), S320. Balen, A.H., Rutherford, A.J., 2007. Management of infertility. BMJ 22, 608611. Bissonnette, F., Cohen, J., Collins, J., et al.; Canadian Fertility and Andrology Society, 2007. Incidence and complications of multiple gestation in Canada: proceedings of an expert meeting. Reprod. Biomed. Online 14, 773790. Bland, J.M., Altman, D.G., 2000. The odds ratio. BMJ 320, 1468. Briggs, A.H., Wonderling, D.E., Mooney, C.Z., 1997. Pulling costeffectiveness analysis up by its bootstraps: a non-parametric approach to condence interval estimation. Health Econ. 6, 327340.

Cost-effectiveness of HP-HMG versus rFSH

British Medical Association, Royal Pharmaceutical Society of Great Britain, 2008. British National Formulary 53. Available from: <http://www.bnf.org/bnf>. Brosens, I., Gordts, S., Valkenburg, M., Puttemans, P., Campo, R., Gordts, S., 2004. Investigation of the infertile couple: when is the appropriate time to explore female infertility? Hum. Reprod. 19, 16891692. Buckett, W., Bentick, B., 1997. The epidemiology of infertility in a rural population. Acta Obstet. Gynecol. Scand. 70, 233237. Cahill, D.J., Wardle, P.G., 2002. Management of infertility. BMJ 325, 2832. Collins, J., 2002. An international survey of the health economics of IVF and ICSI. Hum. Reprod. Update 8, 265277. Commission of the European Communities, 2006. Commission communication. Green paper Confronting demographic change: a new solidarity between the generations. Available from: <http://ec.europa.eu/employment_social/news/2005/ mar/comm200594_en.pdf> (last accessed March 2009). Connolly, M., De Vrieze, K., Ombelet, W., Schneider, D., Currie, C., 2008. A cost per live-birth comparison of HMG and rFSH randomized trials. Reprod. Biomed. Online 17, 756763. Conover, W.J., 1980. Practical Nonparametric Statistics. John Wiley and Sons, New York. Coomarasamy, A., Afnan, M., Cheema, D., van der Veen, F., Bossuyt, P.M., van Wely, M., 2008. 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. 23, 310315. Curtis, L., Netten, A., 2005. Unit Costs of Health and Social Care 2005. Personal Social Services Research Unit, University of Kent, Canterbury. Daya, S., 2003. Pitfalls in the design and analysis of efcacy trials in subfertility. Hum. Reprod. 18, 10051009. Daya, S., Ledger, W., Auray, J.P., et al., 2001. Cost-effectiveness modelling of recombinant FSH versus urinary FSH in assisted reproduction techniques in the UK. Hum. Reprod. 16, 2563 2569. De Vries, M.J., De Sutter, P., Dhont, M., 1999. Prognostic factors in patients continuing in vitro fertilization or intracytoplasmic sperm injection treatment and dropouts. Fertil. Steril. 72, 674 678. El-Toukhy, T., Khalaf, Y., Al-Darazi, K., et al., 2003. Cryo-thawed embryos obtained from conception cycles have double the implantation and pregnancy potential of those from unsuccessful cycles. Hum. Reprod. 18, 13131318. European and Israeli Study Group on Highly Puried Menotropin versus Recombinant Follicle-Stimulating Hormone, 2002. Efcacy and safety of highly puried menotropin versus recombinant follicle-stimulating hormone in in vitro fertilization/ intracytoplasmic sperm injection cycles: a randomized, comparative trial. Fertil. Steril. 78, 520528. European Society of Reproductive Medicine. The ESHRE Capri Workshop Group, 2000. Multiple gestation pregnancy. Hum. Reprod. 15, 18561864. Fenwick, E., OBrien, B.J., Briggs, A., 2004. Cost-effectiveness acceptability curves-facts, fallacies and frequently asked questions. Health Econ. 13, 405415. Filicori, M., Fazleabas, A.T., Huhtaniemi, I., et al., 2005. Novel concepts of human chorionic gonadotrophin: reproductive system interactions and potential in the management of infertility. Fertil. Steril. 84, 275284. Gates, S., Brocklehurst, P., 2004. BJOG. Int. J. Obstet. Gynaecol. 111, 213219. Gold, M.R., Siegel, J.E., Russell, L.B., 1996. Cost-effectiveness in Health and Medicine. Oxford University Press, New York. Gunnell, D.J., Ewings, P., 1994. Infertility prevalence, needs assessment and purchasing. J. Pub. Health 16, 2935.

177
Higgins, J.P.T., Green, S. (Eds.), 2008. Cochrane Handbook for Systematic Reviews of Interventions Version 5.0.1 [updated September 2008]. The Cochrane Collaboration. Available from: <www.cochrane-handbook.org>. Hompes, P.G., Broekmans, F.J., Hoozemans, D.A., Schats, R.; FIRM group, 2008. Effectiveness of highly puried human menopausal gonadotropin vs. recombinant follicle-stimulating hormone in rst-cycle in vitro fertilization-intracytoplasmic sperm injection patients. Fertil. Steril. 89, 16851693. Human Fertilisation and Embryology Authority, 2008. A long term analysis of the Human Fertilisation and Embryology Authority Register data (19912006). Version 1 Revision 4 Extract Date 23/ 01/2008. Jain, T., Harlow, B.L., Hornstein, M.D., 2002. Insurance coverage and outcomes of in vitro fertilization. NEJM 347, 661666. Kennedy, R., Kingsland, C., Rutherford, A., Hamilton, M., Ledger, W., 2006. Implementation of the NICE guideline recommendations from the British Fertility Society for national criteria for NHS funding of assisted conception. Hum. Fertil. 9, 181189. Land, J.A., Evers, J.L., 2003. Risks and complications in assisted reproduction techniques: report of an ESHRE consensus meeting. Hum. Reprod. 18, 455457. Ledger, W.L., Anumba, D., Marlow, N., Thomas, C.M., Wilson, E.C., 2006. The costs to the NHS of multiple births after IVF treatment in the UK. Br. J. Obstet. Gynaecol. 113, 2125. Lloyd, A., Kennedy, R., Hutchinson, J., Sawyer, W., 2003. Economic evaluation of highly puried menotropin compared with recombinant follicle-stimulating hormone in assisted reproduction. Fertil. Steril. 80, 11081113. Lukassen, H.G., Braat, D.D., Wetzels, A.M., et al., 2005. Two cycles with single embryo transfer versus one cycle with double embryo transfer: a randomized controlled trial. Hum. Reprod. 20, 702708. Lunenfeld, B., Van Steirteghem, A., Foundation, Bertarelli., 2004. Infertility in the third millennium: implications for the individual, family and society: condensed meeting report from the Bertarelli Foundations second global conference. Hum. Reprod. Update 10, 317326. Lyerly, A.N., Steinhauser, K., Voils, C., et al., 2010. Fertility patients views about frozen embryo disposition: results of a multi-institutional US survey. Fertil. Steril. 93, 499509. Meldrum, D.R., Silverberg, K.M., Bustillo, M., Stokes, L., 1998. Success rate with repeated cycles of in vitro fertilization embryo transfer. Fertil. Steril. 69, 10051009. Mladovsky, P., Sorenson, C., 2009. Public Financing of IVF: A Review of Policy Rationales. Health Care Anal. Apr 3 (Epub ahead of print). Moustafa, M.K., Sheded, S.A., El Aziz Mousta, M.A., 2008. Elective single embryo transfer versus double embryo transfer in assisted reproduction. Reprod. Biomed. Online 17, 8287. National Institute for Clinical Excellence (NICE), 2004, 2004. Guide to the Methods of Technology Appraisal. NICE, London, UK. Out, H.J., Mannaerts, B.M., Driessen, S.G., Bennink, H.J., 1995. A prospective, randomized, assessor-blind, multicentre study comparing recombinant and urinary follicle stimulating hormone (Puregon versus Metrodin) in in-vitro fertilization. Hum. Reprod. 10, 25342540. Platteau, P., Smitz, J., Albano, C., Srensen, P., Arce, J.C., Devroey, P., 2004. Exogenous luteinizing hormone activity may inuence the treatment outcome in in vitro fertilization but not in intracytoplasmic sperm injection cycles. Fertil. Steril. 81, 14011404. Platteau, P., Andersen, A.N., Balen, A., et al.; Menopur Ovulation Induction (MOI) Study Group, 2006. Similar ovulation rates, but different follicular development with highly puried menotrophin compared with recombinant FSH in WHO Group II anovulatory infertility: a randomized controlled study. Hum. Reprod. 21, 17981804.

178
Platteau, P., Nyboe Andersen, A., Loft, A., Smitz, J., Danglas, P., Devroey, P., 2008. Highly puried HMG versus recombinant FSH for ovarian stimulation in IVF cycles. Reprod. Biomed. Online 17, 190198. Salumets, A., Suikkari, A.M., Makinen, S., Karro, H., Roos, A., Tuuri, T., 2006. Frozen embryo transfers: implications of clinical and embryological factors on the pregnancy outcome. Hum. Reprod. 21, 23682374. Smitz, J., Andersen, A.N., Devroey, P., Arce, J.C., Group, M.E.R.I.T., 2007. Endocrine prole in serum and follicular uid differs after ovarian stimulation with HP-hMG or recombinant FSH in IVF patients. Hum. Reprod. 22, 676687. Staniec, J.F.O., Webb, N.J., 2007. Utilization of infertility services: how much does money matter? Health Serv. Res. 42, 971989. Sykes, D., Out, H.J., Palmer, S.J., van Loon, J., 2001. The costeffectiveness of IVF in the UK: a comparison of three gonadotrophin treatments. Hum. Reprod. 16, 25572562. Templeton, A., Morris, J.K., 1998. Reducing the risk of multiple births by transfer of two embryos after in vitro fertilization. N. Engl. J. Med. 339, 573577. Thurin, A., Hausken, J., Hillensjo, T., et al., 2004. Elective singleembryo transfer versus double embryo transfer in in vitro fertilization. N. Engl. J. Med. 351, 23922402. Van Voorhis, B.J., 2006. Outcomes from assisted reproductive technology. Obstet. Gynecol. 107, 183200. Veleva, Z., Vilska, S., Hyden-Granskog, C., Tiitinen, A., Tapanai nen, J.S., Martikainen, H., 2006. Elective single embryo transfer in women aged 3639 years. Hum. Reprod. 21, 20982102. Wechowski, J., Connolly, M., McEwan, P., Kennedy, R., 2007. An economic evaluation of HP-hMG and rFSH based on a large randomised trial. Reprod. Biomed. Online 15, 500506.

J Wex-Wechowski et al.
Wechowski, J., Connolly, M., Schneider, D., McEwan, P., Kennedy, R., 2009. Cost-saving treatment strategies in in vitro fertilization: a combined economic evaluation of two large randomized clinical trials comparing highly puried human menopausal gonadotrophin and recombinant follicle-stimulating hormone alpha. Fertil. Steril. 91, 10671076. Wolfenson, C., Groisman, J., Couto, A.S., et al., 2005. Batch-tobatch consistency of human-derived gonadotrophin preparations compared with recombinant preparations. Reprod. Biomed. Online 10, 442454. Ziebe, S., Lundin, K., Janssens, R., Nyboe Andersen, A., Arce, J.C., 2006. Cryosurvival and frozen embryo replacement outcome after ovarian stimulation with highly puried menotrophin or recombinant FSH in IVF cycles. Hum. Reprod. 21 (Suppl. 1), i38 i39. Ziebe, S., Lundin, K., Janssens, R., Helmgaard, L., Arce, J.C., IT, M.E.R.MERIT (Menotrophin vs Recombinant FSH in vitro Fertilisation Trial) Group, 2007. Inuence of ovarian stimulation with HP-hMG or recombinant FSH on embryo quality parameters in patients undergoing IVF. Hum. Reprod. 22, 2404 2413. Zwart van Rijkom, J.E., Broekmans, F.J., Leufkens, H.G., 2002. From HMG through puried urinary FSH preparations to recombinant FSH: a substitution study. Hum. Reprod. 17, 857 865. Declaration: JW and AA were provided an unrestricted research grant from Ferring Pharmaceuticals in support of this project. Received 12 November 2009; refereed 18 February 2010; accepted 6 May 2010.