A cost-effectiveness modeling evaluation comparing a biosimilar follitropin alfa preparation with its reference product for live birth outcome in Germany, Italy and Spain

Abstract

Background/objective: Although biosimilar drugs may be cheaper to purchase than reference biological products, they may not be the most cost-effective treatment to achieve a desired outcome. The analysis reported here compared the overall costs to achieve live birth using the reference follitropin alfa (GONAL-f) or a biosimilar (Ovaleap) in Spain, Italy and Germany.

Methods: Patient and treatment data was obtained from published sources; assisted-reproductive technology, gonadotropin, follow-up and adverse-event-related costs were calculated from tariffs and reimbursement frameworks for each country. Incremental cost-effectiveness ratios (ICERs) were calculated from the difference in costs between reference and biosimilar in each country, divided by the difference in live-birth rates. Mean cost per live birth was calculated as total costs divided by the live-birth rate.

Results: The published live birth rates were 32.2% (reference) and 26.8% (biosimilar). Drug costs per patient were higher for the reference recombinant human follicle-stimulating hormone in all three countries, with larger cost differences in Germany (€157.38) and Italy (€141.50) than in Spain (€22.41). The ICER for the reference product compared with the biosimilar was €2917.47 in Germany, €415.43 in Spain and €2623.09 in Italy. However, the overall cost per live birth was higher for the biosimilar in all three countries (Germany €8135.04 vs. €9185.34; Italy €8545.22 vs. €9733.37; Spain €14,859.53 vs. €17,767.19). Uncertainty in efficacy, mean gonadotropin dose and costs did not have a strong effect on the ICERs.

Conclusions: When considering live birth outcomes, treatment with the reference follitropin alfa was more cost effective than treatment with the biosimilar follitropin alfa.

Keywords:

• Cost per live birth
• incremental cost-effectiveness ratio
• ICER
• follitropin alfa
• Spain
• Italy
• Germany
• recombinant human follicle-stimulating hormone

JEL classification codes:

• D40
• D61

Introduction

A biosimilar medicinal product (biosimilar) is a biological product that has been developed to be similar to an authorized biological product (reference product)¹. To obtain regulatory approval, there must be equivalence in the pharmacological, pharmacokinetic, toxicological, efficacy and safety profiles between the reference and biosimilar products¹. These equivalence data is generally obtained from comparability studies, which are not as extensive as the clinical study programs required to gain regulatory approval for the reference product, meaning that the development costs for biosimilars are usually lower than those for the reference¹. Owing to these lower development costs, biosimilars may be marketed at a lower purchase price, providing cost savings on the purchase price compared with the reference product. However, direct comparison of the purchase prices of reference and biosimilar drugs may not be the most appropriate index on which market-access decisions should be based. To illustrate this potential disparity, we consider the comparison of reference and biosimilar recombinant human follicle-stimulating hormone (r-hFSH; follitropin alfa) preparations used in assisted reproductive technologies (ART) treatment.

The most important goal of ART is live birth; however, the endpoint stipulated by the European Medicines Agency (EMA) for comparative trials of recombinant human follitropin alfa preparations is number of oocytes retrieved². Although there is a strong association between the number of oocytes retrieved and live-birth rate³, it would be reasonable to assume that the main objective of treatment – live birth – should be the result on which any considerations of cost-effectiveness are based.

GONAL-f^a first received marketing authorization from the European Medicines Agency (EMA) in 1995⁴, and is the reference product against which biosimilar follitropin alfa containing products are compared^5–7. Ovaleap^b, a biosimilar follitropin alfa, received marketing authorization in Europe in September 2013⁸, based upon a comparability study in which similar numbers of oocytes were retrieved from patients on the different treatments (reference follitropin alfa 12.1 [standard deviation (SD) 6.7] vs. biosimilar follitropin alfa 12.2 [SD 6.7]; difference 0.03 [95% confidence interval -0.76 to 0.82])⁹.

Clinical and on-going pregnancy rates were also reported for the two preparations (clinical pregnancy rates: biosimilar follitropin alfa 28.1% [43/153] and reference follitropin alfa 35.6% [52/146]; on-going pregnancy rates per patient: biosimilar follitropin alfa 27.5% [42/153] and reference follitropin alfa 33.6% [49/146]). A total of 89.1% (41/46) of patients in the biosimilar follitropin alfa group and 88.7% (47/53) in the reference follitropin alfa group went on to have a live birth⁹. Based on a comparability study of biosimilar follitropin alfa versus reference follitropin alfa for the stimulation of follicular development, similar numbers of oocytes were retrieved from patients on either treatment (difference in oocytes retrieved 0.03 [95% confidence interval -0.76 to 0.82])⁹ and similar pregnancy rates per embryo transfer were also reported in the first and second cycles (biosimilar follitropin alfa 40.2% and 38.5%, respectively; reference follitropin alfa 48.2% and 27.8%, respectively)⁹. However, differences in the cost to achieve live birth between the two treatments have not been investigated.

To analyze the difference in the overall cost to achieve a live birth with either reference or biosimilar follitropin alfa, a cost-effectiveness modeling analysis was conducted using cost data from three countries: Germany (where biosimilar follitropin alfa costs less than reference follitropin alfa and all associated costs are obtained separately), Italy (where biosimilar follitropin alfa costs less than reference follitropin alfa and the associated costs are according to a tariff) and Spain (where reference follitropin alfa and biosimilar follitropin alfa cost the same and the associated costs are according to a tariff).

Methods

Clinical data

The cost-effectiveness analysis was based on clinical data from a single phase III study⁹ and the EMA Assessment Report for biosimilar follitropin alfa¹⁰. Data on patient numbers, total follitropin alfa dose, clinical pregnancy rate per embryo transfer, live-birth rate (irrespective of the number of children born) and incidence and severity of ovarian hyperstimulation syndrome (OHSS) was used in the cost-effectiveness model (Table 1). The proportion of patients with infertility due to male factor was similar in each group (biosimilar follitropin alfa 56.2% vs. reference follitropin alfa 52.7%)⁹.

Table 1. Clinical data.

	Reference follitropin alfa	Biosimilar follitropin alfa
	(n = 146)	(n = 153)
Total follitropin alfa dose (IU), mean (SD)	1614 (485)	1536 (496)
Clinical pregnancy rate per embryo transfer, n/N (%)	52/134 (38.8)	43/141 (30.5)
Clinical pregnancy rate per cycle started, n/N (%)	52/146 (35.6)	43/153 (28.1)
Patients with live-born children, n (%)^a	47/146 (32.2)	41/153 (26.8)
OHSS
Overall, n (%)	4 (2.7)	7 (4.6)
Mild/moderate (%)	3 (2.0)	6 (3.9)
Severe (%)	1 (0.7)	1 (0.7)

Abbreviations. OHSS, ovarian hyperstimulation syndrome; SD, standard deviation.

Clinical data was obtained from a phase III study⁹ and the EMA Assessment report for Ovaleap¹⁰.

^aLive birth is defined as the proportion of randomized patients who had a live birth following treatment with either of the preparations. Multiple births occurred in 22 women, or 25% of the participants who had live births. In the biosimilar follitropin alfa group, 13/41 patients who had live births (31.7%) had twins. There were 8 patients with twins and 1 patient with triplets in the reference follitropin alfa group (9/47 live births [19.1%]).

Costs per patient

Germany

The costs for all ART-related procedures, drugs and OHSS treatment were obtained separately (Table 2), rather than as part of an overall tariff. Ex-pharmacy drug prices were obtained from the Lauer Taxe¹¹ and represent prices for June 2017. The cost for follitropin alfa was calculated from the most economical combination of packages to cover at least the mean total dose per patient (one package each of 300 IU, 450 IU and 900 IU); excess follitropin alfa was treated as waste. Procedure-related costs were obtained from the Einheitlichen Bewertungsmasstab (EBM) online catalogue¹². Costs related to OHSS were obtained from the Fallpauschalen-Katalog¹³ 2017 and the GKV-Spitzenverband¹⁴. These were calculated from diagnosis-related group (DRG) tariffs, using the base tariff for 2017 (€3376.11) and the valuation ratio for DRG X62Z (0.563).

Table 2. Cost data (Germany).

Routine Intervention	Cost per unit (€)
Ex-pharmacy price for 300, 450 and 900 IU (reference follitropin alfa)	186.41/274.13/537.26^11
Ex-pharmacy price for 300, 450 and 900 IU (biosimilar follitropin alfa)	150.00/220.00/430.01^11
Ex-pharmacy price hCG: Ovitrelle 6500 IU	52.97^11
Ex-pharmacy price GnRHa: Metrelef N2 package	162.70^11
Ex-pharmacy price Progestan and Crinone ×0.5	58.20^11
Clinical follow-up visits until first ultrasound: gynecological visit + ultrasound (general rate and treatment plan)	22.33^12
Intracytoplasmic sperm injection consultation for couples only	8.53^12
Preparation of sperm	17.69^12
Follicle puncture	54.97^12
Outpatient surgery	47.07^12
In-vitro fertilization and embryo transfer	923.69^12
In-vitro fertilization with no embryo transfer	787.54^12
Discontinuation before ovum pick-up	209.65^12
Intracytoplasmic sperm injection and embryo transfer	1301.30^12
Intracytoplasmic sperm injection with no embryo transfer	1165.36^12
Follow-up blood test for β-HCG (pregnancy test)	6.10^12
OHSS mild/moderate
N.2 Hematology laboratory test	0.25^12
N.2 Hemoatocrit laboratory test	0.25^12
N.2 Creatinine laboratory test	8.45^12
N.2 Electrolytes laboratory test	1.00^12
N.2 Hepatic laboratory tests	10.25^12
OHSS severe
DRG X62Z “Poisoning/toxic effects of drugs, medications and all other substances or consequences of a medical treatment or certain frostbites and other traumas”	1900.75^13,Citation14

Abbreviations. ART, assisted reproductive technologies; β-HCG, beta human chorionic gonadotropin; DRG, diagnosis-related group; EBM, Einheitlichen Bewertungsmassstab; GKV, Geschäftsverkehr; IU, international units; GnRHa, gonadotropin releasing hormone agonist; OHSS, ovarian hyperstimulation syndrome.

For Germany, costs for all treatments associated with ART, drug costs and costs for OHSS treatment were obtained separately. The cost of drugs were taken from the Lauer Taxe¹¹ and represents the status as of June 2017. The German analysis considered EBM tariffs, which were retrieved from the EBM online catalogue¹², whereas the DRG tariffs were calculated with the base tariff for 2017 (€3376.11) and the valuation ratio for DRG X62Z (0.563) taken from the Fallpauschalen-Katalog 2017¹³ and the GKV Spitzenverband¹⁴.

Italy and Spain

Costs were based on DRG tariff 359, which covers all ART procedures, excluding gonadotropin drugs, follow-up visits and adverse events, the costs of which were obtained separately (Table 3). The costs for DRG tariff 359 and gonadotropin drugs (calculated as unitary follitropin alfa costs multiplied by mean total dose) were applied to each patient entering the model for each treatment arm. The cost of the β human chorionic gonadotropin blood test (to confirm pregnancy) was included, and clinical follow-up visit costs were only applied to women who got pregnant. For Italy, the cost of DRG tariff 359 was retrieved from the Tariffa Unica Convenzionale (TUC) 2013 price list¹⁵ and drug costs were obtained from the Gazzetta Ufficiale Elenco Atti dell’emittente Agenzia Italiana del Farmaco¹⁶. For Spain, the cost of DRG tariff 359 was retrieved from the Registro de Atlas de los Hospitales Generales del Sistema Nacional de Salud¹⁷, and drug costs were obtained from the eSalud Sanitary Costs database¹⁸.

Table 3. Cost data (Italy and Spain).

Routine Intervention	Cost per unit (€)
	Italy	Spain
DRG 359 “Uterus and adnexa non-malignant, without cost specification”	2124.52^15	4289.51^17
Ex-factory price reference follitropin alfa (75 IU)	27.38^16	20.16^18
Ex-factory price biosimilar follitropin alfa (75 IU)	21.90^16	20.16^18
Clinical follow-up visits until first ultrasound: gynecological visit and ultrasound	52.68^15	83.07^17
Follow-up blood test for β-HCG (pregnancy test)	9.98^15	15.00^17
OHSS mild/moderate
N.2 Gynecology visit	20.66^15	50.90^17
N.2 Ultrasound of lower abdomen	32.02^15	32.17^17
N.2 Hematology laboratory test	7.41^15	10.60^17
N.2 Hematocrit laboratory test	0.69^15	6.11^17
N.2 Creatinine laboratory test	1.13^15	2.90^17
N.2 Electrolytes laboratory test	4.30^15	3.86^17
N.2 Hepatic laboratory tests	14.30^15	14.50^17
OHSS severe
DRG 369 “Menstrual disorders and other disorders of the female reproductive system”	800.35^15	1657.28^17

Abbreviations. ART, assisted reproductive technologies; β-HCG, beta human chorionic gonadotropin; DRG, diagnosis-related group; IU, international units; OHSS, ovarian hyperstimulation syndrome.

For Italy and Spain, this analysis considered the DRG 359 tariff, which covers all ART procedures, excluding gonadotropin drugs, follow-up visits and adverse event costs, which have been cited separately. The Tariff was retrieved from the Tariffa Unica Convenzionale 2013¹⁵ pricelist from the Italian Ministry of Health website and from the Registro de Atlas de los Hospitales Generales del Sistema Nacional de Salud¹⁷, for the Italian and Spanish costs respectively. Drug costs were retrieved from the Gazzetta Ufficiale Elenco Atti dell’emittente Agenzia Italiana del Farmaco¹⁶ and the eSalud Sanitary Costs database¹⁸ for the Italian and Spanish costs respectively.

Cost effectiveness

Cost effectiveness is reported as the incremental cost-effectiveness ratio (ICER), which is calculated by dividing the difference in cost by the difference in efficacy. The ICER shows the additional cost to achieve an incremental increase in efficacy, enabling direct comparison of the cost effectiveness between countries.

Modeling the cost per live birth

A decision-tree model was developed using Microsoft Excel Professional 2016. All costs per patient were applied to the clinical data, simulating the cost per live birth in each country. OHSS costs were estimated for all three countries according to the probability of experiencing OHSS multiplied by unitary costs. Total cost per live birth was calculated by summing up all costs and then dividing this value by the efficacy parameter (live-birth rate) to obtain the mean cost per live birth in each country.

Sensitivity analyses

Univariate sensitivity analyses considering the effects of efficacy, gonadotropin cost and mean gonadotropin dose were performed to investigate the effect (lowest and highest limits) of each of these variables on the ICERs for each country.

Results

Clinical data

The live-birth rates were 32.2% for reference follitropin alfa and 26.8% for biosimilar follitropin alfa (Table 4). These rates were reported as the proportion of randomized patients in the equivalence trial of reference follitropin alfa versus biosimilar follitropin alfa who had a live birth⁹.

Table 4. Cost-effectiveness analysis.

	Total cost per patient (€)	Incremental cost (€)	Efficacy	Incremental efficacy*	ICER (€)	Cost per live birth (€)
Germany
Biosimilar follitropin alfa	2461.43		0.267973			9185.34
Reference follitropin alfa	2618.81	157.38	0.321917	0.053944	2917.47	8135.04
Italy
Biosimilar follitropin alfa	2609.36		0.267973			9737.37
Reference follitropin alfa	2750.86	141.50	0.321917	0.053944	2623.09	8545.22
Spain
Biosimilar follitropin alfa	4761.14		0.267973			17,767.19
Reference follitropin alfa	4783.55	22.41	0.321917	0.053944	415.43	14,859.53

Abbreviation. ICER, incremental cost-effectiveness ratio.

Costs per patient

Owing to the differences in the cost of the drugs, the per-patient total-cost estimates were higher for reference follitropin alfa than for biosimilar follitropin alfa in all three countries. The difference was larger in Germany (€157.38) and Italy (€141.50) than in Spain (€22.41). Overall, the costs for ART-related procedures and drugs in Germany and Italy were approximately half the costs of those in Spain, irrespective of which treatment was used.

Cost effectiveness

According to the ICER values, the incremental increase in efficacy (live-birth rate) from 26.8% with biosimilar follitropin alfa to 32.2% with reference follitropin alfa would incur an additional cost (best estimate) when using reference follitropin alfa of €2917.47 in Germany, €2623.09 in Italy and €415.43 in Spain (Table 4).

Modeling the costs per live birth

The estimated overall costs per live birth per patient per cycle were lower in all three countries for patients treated with reference follitropin alfa compared with those treated with biosimilar follitropin alfa (difference in Germany €1050.30; difference in Italy €1192.15; difference in Spain €2907.66 [Table 4]).

Sensitivity analyses

In the univariate sensitivity analyses, uncertainty in the parameters (efficacy, gonadotropin cost and mean gonadotropin dose) did not have a strong effect on the ICER values in any of the countries (Figure 1). As the parameters were varied from their lower to their upper limits (±5%), the ICERs changed from approximately €2000 to €4300 for Germany, from approximately €2000 to €3700 for Italy and from approximately €10 to €800 for Spain. Owing to the fixed combination of drug packages in the model for Germany, the lower ICER limit for mean dose was the same as the ICER. The minimum incremental dose delivered from the different package combinations in the German model was 150 IU; therefore, a 5% reduction in the mean dose would not result in an IU value under the threshold for a more economical combination of packages (1500 IU).

Figure 1. Sensitivity analyses for effect of changes in efficacy, mean gonadotropin dose and gonadotropin cost on incremental cost-effectiveness ratios in: A) Germany, B) Italy and C) Spain.

Discussion

Many new-to-market interventions, including biosimilars, provide equivalent efficacy to existing treatments at a lower purchase price and may, at first, seem to provide cost savings. Our aim was to compare the overall costs to achieve a live birth, the goal of infertility treatment, with reference or biosimilar follitropin alfa in three countries. We showed that despite cost prices and ICERs that favored the use of biosimilar follitropin alfa, the overall cost incurred to achieve a live birth was actually lower for reference follitropin alfa in all three countries.

The ICER is a useful metric to compare the cost-effectiveness of treatment options (either related or unrelated) for attaining specific clinical outcomes. The results from ICER analyses, used in conjunction with willingness-to-pay thresholds, are often instrumental when making a decision on how to allocate limited healthcare payer resources. This enables the exclusion of treatments or interventions that are not considered cost-effective according to pre-specified thresholds. In the analysis reported here, the incremental cost between the two drugs shows that biosimilar follitropin alfa is cheaper to purchase than reference follitropin alfa; however, owing to the increased live birth rate observed after ovarian stimulation with reference follitropin alfa (32.2%) when compared with ovarian stimulation with biosimilar follitropin alfa (26.8%), the actual cost per live birth was lower with reference follitropin alfa compared with biosimilar follitropin alfa. In addition to this favorable outcome, the ICER results for all countries were well below internationally used thresholds. Owing to these findings, and despite its higher purchase price, reference follitropin alfa is a cost-effective treatment to achieve the desired outcome of pregnancy compared with the biosimilar follitropin alfa in all countries assessed. We show that, despite a cheaper purchase price or favorable ICER, the newer treatment may not be the most cost-effective option to achieve a desired efficacy outcome (live birth).

The sensitivity analyses conducted to investigate the impact of variations in efficacy, gonadotropin cost and mean gonadotropin dose demonstrated that these parameters did not have a strong overall effect on the ICER values, highlighting the robustness of the model. However, despite the absence of an overall effect, the sensitivity analyses demonstrated that there were country-specific differences in the parameters that had the greatest effects on ICERs. In Germany all three parameters had an effect on ICERs, possibly because of a lack of flexibility in the flat rate used in the analyses. In Italy, variations in efficacy had the greatest effect; and in Spain variations in gonadotropin cost and mean gonadotropin dose had the greatest effects on ICER.

The results of this analysis comparing reference follitropin alfa with biosimilar follitropin alfa are consistent with the results of a previous analysis comparing the cost-effectiveness of reference follitropin alfa and a different biosimilar follitropin alfa (Bemfola^c) to achieve live birth in two countries (Italy and Spain)¹⁹. In this previous analysis, the ICER for reference follitropin alfa compared with the biosimilar follitropin alfa was €3600 in Italy and €900 in Spain, whereas the estimated overall cost per live birth was €367 less with reference follitropin alfa than with the biosimilar follitropin alfa in Italy and €1211 less with reference follitropin alfa than with the biosimilar follitropin alfa in Spain. The results from this previous analysis and the results reported here make a strong case that any perceived cost savings from the lower purchase price of a biosimilar follitropin alfa do not reflect the true costs incurred to achieve live birth.

Some potential limitations should be taken into account when interpreting the results of this analysis. Firstly, the primary outcome of the study from which the clinical data for the model was extracted was the number of oocytes retrieved, rather than live-birth rate. This study was, therefore, not powered to analyze differences in live birth rate. Furthermore, although the source study was a multinational study of women undergoing ART treatment that was carried out across several centers, the overall sample size was relatively small (around 150 patients per group), and only 41 patients receiving biosimilar follitropin alfa and 47 patients receiving reference follitropin alfa had a live birth. While we would have preferred to have had a more robust dataset on which to base our calculations, biosimilar follitropin alfa is still a new product and, consequently, there is little published data available for use. Despite this, our analysis still provides valuable information on the costs and benefits of the two preparations. Secondly, as is inherent with all modeling studies, this analysis includes assumptions and biases that are not based on real-world experience. Furthermore, any inaccuracies and imprecisions as a result of these assumptions and biases have the potential to be compounded throughout subsequent calculations. It should also be noted that ex-factory prices were used in Italy and Spain, whereas ex-pharmacy prices were used in Germany, and the model does not account for dose fluctuations or drug wastage in any of the regions studied. These were not included in the model as no combination of pen injectors could be used to exactly deliver the total mean doses used in the model. This may be of particular relevance for the estimated drug costs for Germany, as costs were based on the most commonly purchased pack sizes (300, 450 and 900 IU) for one treatment cycle, leading to a high degree of wastage for biosimilar follitropin alfa. Thirdly, the physical costs included in the model are not the only treatment-related and adverse-event-related costs, and additional costs that could also be considered include societal costs relating to adverse events and treatment burden, as well as the psychological impact of having to repeat a treatment cycle following the failure of a cycle. However, these costs would need to be applied subjectively, and it would be very difficult to correctly estimate them in each country. Fourthly, the study from which the clinical data was sourced was conducted in a selected group of women (aged <37 years, healthy, and excluding those at risk for OHSS, those with previous abortions, >2 previous IVFs and poor ovarian response; furthermore, the dose of follitropin alfa administered was fixed and adjusted to obtain the requisite number of follicles). Accordingly, the results of our study may not be generalizable for all populations of women undergoing ART treatment. Finally, at present there are no nationally or internationally agreed ICER thresholds for live birth from ART treatment that can be used to put the results of the analysis into the context of the likelihood of a treatment being adopted in each of the countries investigated.

Conclusions

The results of this analysis show that, compared with biosimilar follitropin alfa, treatment with reference follitropin alfa is more cost-effective to achieve live birth. This applies when the biosimilar has a lower purchase price than the reference product and also when the biosimilar and reference products have similar costs but the country-specific pricing frameworks for calculating the associated costs are different.

Transparency

Declaration of funding

This analysis was funded by Merck KGaA, Darmstadt, Germany.

Author contributions: S.G. contributed to the data evaluation and the critical revision of the manuscript. M.F. and K.B. contributed to the design, inception and performance of the study, and the critical revision of the manuscript. M.L. contributed to the design of the study and the critical revision of the manuscript. C.R. and N.C. designed and developed the models for Italy and Spain, retrieved the data for the analysis, analyzed the results and contributed to the critical revision of the manuscript.

Declaration of financial/other relationships

S.G. and C.R. have disclosed that they have no significant relationships with or financial interests in any commercial companies related to this study or article. M.F. has disclosed that he has previously received funding from Merck KGaA, Darmstadt, Germany. M.L. has disclosed that she is an employee of Merck Serono SpA, an affiliate of Merck KGaA, Darmstadt, Germany. N.C. has disclosed that she is an employee of Quintiles IMS, Milan, Italy. K.B. has disclosed that he has received honoraria from Merck KGaA, Takeda and Stiftung Endometriose Forschung.

A peer reviewer on this manuscript received research grants from Merck. The remaining peer reviewers have no relevant financial or other relationships to disclose.

Medical writing support was provided by Steven Goodrick of inScience Communications, Springer Healthcare, UK, and funded by Merck KGaA, Darmstadt, Germany.

Acknowledgements

The authors would like to thank Dirk Eheberg for his contribution to the cost-effective analysis.

Notes

Notes

a GONAL-f is a registered trade name of Merck KGaA, Darmstadt, Germany

b Ovaleap is a registered trade name of Teva, Petah Tikva, Israel

c Bemfola is a registered trade name of Gedeon Richter UK Ltd, Budapest, Hungary