2013

http://informahealthcare.com/gyeISSN: 0951-3590 (print), 1473-0766 (electronic)

Gynecol Endocrinol, 2013; 29(6): 520–529! 2013 Informa UK Ltd. DOI: 10.3109/09513590.2012.758703

HP-HUMAN FSH VERSUS RFSH IN IVF-ICSI CYCLES

Clinical efficacy and cost-effectiveness of HP-human FSH (Fostimon�)versus rFSH (Gonal-F�) in IVF-ICSI cycles: a meta-analysis

Sandro Gerli1, Vittorio Bini2, Alessandro Favilli1, and Gian Carlo Di Renzo1

1Department of Obstetrics and Gynecology, University of Perugia, Perugia, Italy, and 2Department of Internal Medicine, University of Perugia,

Perugia, Italy

Abstract

Clinical efficacy of human-derived follicle-stimulating hormone (FSH) versus recombinant FSH(rFSH) in IVF-ICSI cycles has long been compared, but no clear evidence of the superiority of apreparation over the other has been found. Human gonadotropins have been often groupedtogether, but a different glycosylation may be present in each preparation, thereforeinfluencing the specific bioactivity. To exclude confounding factors, a meta-analysis and a cost-effectiveness analysis were designed to compare effectiveness and cost-effectiveness of aspecific highly purified human FSH (HP-hFSH) (Fostimon�) versus rFSH (Gonal-F�) in IVF/ICSIcycles. Research methodology filters were applied in MEDLINE, Current Contents and Web ofScience from 1980 to February 2012. Eight randomized trials met selection criteria. The meta-analysis showed no significant differences between rFSH and HP-hFSH treatment in live-birthrate (odds ratio [OR] 0.84, 95% confidence interval [CI] 0.63–1.11), clinical pregnancy rate (OR0.85, 95% CI 0.68–1.07), number of oocytes retrieved, number of mature oocytes and days ofstimulation. The cost-effectiveness ratio was E7174 in the rFSH group and E2056 in the HP-hFSH group. HP-hFSH is as effective as rFSH in ovarian stimulation for IVF-ICSI cycles, but thehuman preparation is more cost-effective.

Keywords

Cost-effectiveness, effectiveness, FSH, IVF,meta-analysis, outcome

History

Received 2 November 2012Accepted 4 December 2012Published online 28 January 2013

Introduction

Recombinant follicle-stimulating hormone (rFSH) was introducedin the market as a significant technological and clinicalimprovement [1], but currently there is no clear evidencedemonstrating the superiority of rFSH over human gonadotropinssuch as human follicle-stimulating hormone (hFSH) and humanmenopausal gonadotropins (hMG) in terms of effectiveness.

Many contradicting results have been reported during the last15 years: in terms of clinical efficacy, there are a number of meta-analyses demonstrating no significant difference in clinical/ongoing pregnancy/live-birth rate, miscarriage rate, multiplepregnancy rate and incidence of ovarian hyperstimulationsyndrome (OHSS) between rFSH and hMG [2,3] and betweenrFSH and hFSH [4]. A systematic review and meta-analysisshowed a significant 4% increase in live-birth rate with the use ofhMG when compared with rFSH following a long down-regulation protocol in IVF-ICSI treatment cycles [5]. A recentCochrane review compared rFSH with the three main types ofhuman gonadotropins: hMG, hFSH and highly purified hFSH(HP-hFSH) for ovarian stimulation in women undergoing IVF orICSI treatment cycles, demonstrating an equal effectiveness inpregnancy outcomes and a small or no difference in live births, inOHSS in any of the down-regulation groups [6].

Analyses of human gonadotropins have revealed importantdifferences in terms of FSH and luteinizing hormone (LH)

content, protein contamination and polymorphism of theseglycoproteins that became available in the last 20 years. TheFSH isoforms are characterized by heterogeneity of differingamounts of sialic acid and complexity of the attached oligosac-charide structures both of which appear to influence the specificbioactivity [7]: rFSH presents a lower grade of carbohydratebranched structures and less-acidic isoforms than hFSH [8,9].

Regarding the hFSH preparation, the steps involved in thepurification of hFSH products can affect the FSH isoformcomposition: it is therefore erroneous to address clinical andendocrinological issues with results of studies, where hFSH isgrouped together with HP-hFSH [10] or even more when rFSH orhFSH are compared to hMG [5], with a profound difference ingonadotropin content.

Moreover, even if the same subgroup is considered (HP-hFSH),different FSH isoforms are included in different preparations: forinstance Metrodin (Metrodin HP�; Serono International SAGeneva, Switzerland) cannot be considered as equivalent asFostimon (Fostimon�; IBSA Institute Biochimique SA Lugano,Switzerland), although both preparations are characterized as‘‘HP-hFSH’’. A different glycosylation may be present in eachpreparation, being higher in Fostimon�, with a specific bioactivityvery close to the range secreted by the pituitary gland [11].

A new systematic review and meta-analysis of 22 randomizedcontrolled trials (RCTs) demonstrates the higher efficiency ofFostimon� compared to Metrodin-HP� in primary clinicaloutcomes such as live-birth/ongoing pregnancy rate [12].

To exclude confounding factors which may alter the results, itis of utmost importance to compare the recombinant FSH with themost effective human preparation with similar gonadotropin content.

Address for correspondence: Sandro Gerli, MD, Department of Obstetricsand Gynecology, University of Perugia, Ospedale Santa Maria dellaMisericordia, 06132 Perugia, Italy. Tel: þ39 075 5783231. Fax:þ39 075 5783829. E-mail: gerber@unipg.it

Gyn

ecol

End

ocri

nol D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y U

nive

rsity

of

Cal

gary

on

10/0

4/13

For

pers

onal

use

onl

y.

In order to explore the effectiveness of the different FSHisoform composition, we decided to compare by means of a meta-analysis the rFSH (Gonal-F�; Serono International SA Geneva,Switzerland) with a single preparation of HP-hFSH (Fostimon�;IBSA Institute Biochimique SA Lugano, Switzerland) in IVF-ICSI cycles.

Second, we elaborated a cost-effectiveness analysis of theconsidered studies to include a financial perspective which is amajor outcome in our practice.

Methods

Identification and eligibility of studies

On July 2012, we performed a systematic search with neitherlanguage nor time limits of articles published in peer-reviewedhealthcare-related journals to identify pertinent studies. Researchmethodology filters were applied in MEDLINE [13], CurrentContents and Web of Science from 1980 to July 2012. We usedthe following keywords: gonadotropins, menotropins, FSH,recombinant FSH, urinary FSH and IVF. No additional articleswere identified through the references cited in the first series ofarticles selected.

Study selection, assessment of study quality and dataextraction

Nine studies were selected with any cause of infertility, with noobvious overlap of subjects with other studies [14–22]. Thetherapeutic interventions were r-FSH (Gonal-F�) or HP-hFSH(Fostimon�) for ovarian stimulation in IVF or ICSI treatment,following a long down-regulation protocol with GnRH agonist. Inone study [15] metformin was given in both groups and thiscontemporaneous administration was not considered with asignificant impact for the results of the study (Table 1).

Methodological quality of the trials was assessed using some ofthe parameters proposed by Jadad et al. [23]. This validated scorewas in the range of 0–5. Studies were scored according to thepresence of three key methodological features of randomization,blinding and accountability of all patients, including withdrawals.A trial was excluded due to the non-RCT nature of the study [19].Eight RCTs fulfilled our selection criteria and were included in themeta-analysis [14–22]. One study [20] was previously included in amulticenter trial [16] and most of results were excluded from theanalysis. Only data of live births (not available from themulticenter trial) were recorded and included in the meta-analysis.In one trial [22] three groups of patients were randomized, but onlytwo groups were selected (patients only treated with hFSH andrFSH). Four trials [15–17,20] were sponsored by IBSA, InstituteBiochimique SA Lugano, Switzerland, four trials were not funded[14,18,21,22]. The review included a total of 1437 patients.

Key data were extracted independently by two authors (SG &VB). For each study, we extracted the following characteristics:authors, journal and year of publication; study design, randomiza-tion and concealment, characteristics of participants; inclusionand exclusion criteria; interventions such as mean type of long-stimulation protocol, FSH starting dose, IVF or ICSI procedures;primary and secondary outcomes such as mean number (�SD)and quality of retrieved oocytes, clinical pregnancy rate, live-birthrate, number of days of FSH stimulation, doses of FSH andstimulation duration. If any data for trial reports were missing(e.g. out of scope of the study endpoints) the investigator of thetrial was contacted by e-mail for additional information.

Statistical analysis

In meta-analysis of binary data we evaluated the OR with 95%confidence intervals (CI). Meta-analysis of continuous data was

performed evaluating weighted mean difference (WMD). Datawere extracted from selected studies and combined applying arandom-effects model (DerSimonian and Laird method) [24],which incorporates heterogeneity of effects. Statistical hetero-geneity across trials was assessed by the Cochran’s Q test and I2

statistics, which describe the percentage of total variation acrossstudies that is due to heterogeneity rather than chance [25].Heterogeneity was considered significant if p50.10 and I2 wasmore than 50%. Additionally we assessed the probability ofpublication bias with Begg’s and Egger’s tests [26–28]. In asensitivity analysis, we formally assessed the influence ofindividual studies on pooled estimates by excluding one studyat the time according to Tobias’ method [29]. If the point estimatewith one study omitted lay outside the 95%CI of the overallestimate with all available trials contributing, the study wasindicated as having excessive influence. Meta-analysis calcula-tions were performed using StatsDirect statistical software version2.7.2 (StatsDirect Ltd, Cheshire, England). The primary outcomesmeasure were OR of live-birth rate and OR of clinical pregnancyrate per randomized woman. Secondary outcome measures weremean number (�SD) and maturity of retrieved oocytes,implantation rate, days of stimulation per randomized woman.In the graphical display of the meta-analyses, a benefit from HP-hFSH would be displayed graphically to the left of the centerlineand a benefit from rFSH would be displayed graphically to theright of the centerline.

Cost-effectiveness analysis

For the present cost-effectiveness analysis we used the dataresulting from the studies. The trials were conducted in Italy infour studies [14,18,21,22], in the USA in one [17], in Egypt in one[15], in France and Hungary in one [16] and in Hungary inone [20]. The cost of a single ampoule of HP-hFSH and rFSH wasdetermined at the time of the study (2011 Italian Formulary) and acost of E0.24 and E0.54 per IU, respectively, was found (1:2.25ratio). We decided to carry out the analysis of cost effectivenessthrough Italian costs for two reasons: (1) the absence of HP-hFSH(Fostimon�) in the U.S. market and (2) the similar cost ratiobetween HP-hFSH (Fostimon�) and r-FSH (Gonal-F�) existing inFrance (E0.22 and E0.42 for hFSH and rFSH, respectively, with aratio of 1:1.90), in Hungary (E0.21 and E0.45 for hFSH andrFSH, respectively, with a ratio of 1:2.14) and in Egypt (55 EGPand 145 EGP¼E0.09 and E0.23 for hFSH and rFSH,respectively, with a ratio of 1:2.55) which are similar to theItalian ratio.

The average cost per cycle was calculated from individualdata, multiplying the cost of a single IU by the weighted means ofIU used per cycle, and cost-effectiveness ratios (CE) werecalculated as the average cost per patient divided by the overallpregnancy rate. The incremental cost-effectiveness ratio (ICER),i.e. the extra cost to achieve one more pregnancy, was calculatedas: ICER¼ (CR�CH)/(ER�EH), where C is cost (of alltreatments of each group of patients), E is effectiveness(pregnancy rate of each group of patients) and the subscripts Rand H denote recombinant and human-derived FSH treatment,respectively. Furthermore, to investigate robustness of the CE todifferent scenarios, we used one-way sensitivity analyses,calculating the CE using a range of estimates of pregnancy ratefrom 0 to 1 and for a range on cost per cycle of rFSH from E200to E2400. We built a decision tree model of a complete cycle inwhich we used all the probabilities of IVF/ICSI: no pregnancies,pregnancies, deliveries and miscarriages observed in both groups.This was similar to the model we adopted for the prospectiverandomized study previously published [30,31]. Decision treeconstruction and sensitivity analyses were performed using

DOI: 10.3109/09513590.2012.758703 rFSH versus hFSH: meta-analysis 521

Gyn

ecol

End

ocri

nol D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y U

nive

rsity

of

Cal

gary

on

10/0

4/13

For

pers

onal

use

onl

y.

Tab

le1

.C

har

acte

rist

ics

of

ran

do

miz

edtr

ials

of

HP

-hF

SH

ver

sus

rFS

Hin

IVF

-IC

SI

cycl

es.

Au

tho

rs,

jou

rnal

,yea

ro

fp

ub

lica

tio

nS

tud

yty

pe

and

des

ign

Par

tici

pan

ts,

incl

usi

on

crit

eria

Excl

usi

on

crit

eria

Inte

rven

tio

ns

Pri

mar

yo

utc

om

esS

eco

nd

ary

ou

tco

mes

Fin

anci

alsu

pp

ort

by

ph

arm

aco

mp

any

Sel

man

HA

etal

.,F

erti

lS

teri

l,2

00

2In

terv

enti

on

alp

aral

lel

ran

do

miz

edo

pen

-la

bel

stu

dy

Co

mp

ute

r-gen

erat

edal

loca

tio

nco

nce

aled

(sea

led

envel

op

es)

Wo

men

un

der

go

ing

afi

rst

IVF

/IC

SI

cycl

e(n¼

26

7)

Age

18

to3

8yea

rs;

BM

I2

0–

26

kg

/m2

infe

rtil

ity

du

eto

tub

alfa

cto

r,m

ale

fact

or

or

un

exp

lain

edin

fert

ilit

y

Gy

nec

olo

gic

abn

orm

alit

ies

or

dis

ease

;p

re-

vio

us

po

or

resp

on

seto

go

na-

do

tro

ph

ins

use

dfo

rIU

I;h

isto

ryo

fse

ver

eO

HS

S;

PC

OS

;az

oo

sper

mia

or

clin

ical

sig

ns

of

infe

ctio

nd

etec

ted

inse

men

anal

ysi

sw

ith

in1

2m

on

ths

bef

ore

the

trea

tmen

t

Lo

ng

dow

n-r

egu

lati

on

wit

hG

nR

Ha

(Tri

pto

reli

n3

.75

mg

i.m

.o

nd

ay2

1o

fp

rev

iou

scy

cle)

Inte

rven

tio

n:

rFS

H(n¼

13

3)

or

HP

-h

FS

H(n¼

13

1)

Sta

rtin

gd

ose

:2

25

IUd

aily

;th

enad

just

ed

10

00

0IU

hC

Gw

hen

�2

foll

icle

so

f�

18

mm

.O

ocy

tes

retr

ieved

34

to3

6h

afte

rh

CG

adm

inis

-tr

atio

n.

Em

bry

os

tran

sfer

red

on

day

3.

Lu

teal

ph

ase

sup

po

rt:

Pro

ges

tero

ne

IM5

0m

gd

aily

No

.o

fo

ocy

tes

mat

ure

Em

bry

oqu

alit

y

Pre

gn

ancy

rate

Imp

lan

tati

on

rate

To

tal

do

seo

fF

SH

adm

inis

tere

d

To

tal

nu

mb

ero

fd

ays

of

stim

ula

tio

n

Fer

tili

zati

on

rate

Em

bry

ocl

eavag

era

te

Liv

e-b

irth

rate

Mis

carr

iag

era

te

Can

cell

atio

nra

te

Inci

den

ceo

fm

od

er-

ate

or

sever

eO

HS

S

No

Mo

ham

edM

Aet

al.,

Fer

til

Ste

ril,

20

06

Inte

rven

tio

nal

ran

do

miz

ed,

op

en-l

abel

,p

aral

lel-

gro

up

con

-tr

oll

edst

ud

y

Co

mp

ute

r-gen

erat

edal

loca

tio

nn

ot

stat

ed

Wo

men

un

der

go

ing

firs

tIV

Fcy

cle

du

rin

gth

est

ud

yp

erio

d(n¼

25

7)

Age4

39

yrs

PC

OS

;F

SH

41

0IU

/mL

Lo

ng

dow

n-r

egu

lati

on

wit

hG

nR

Ha

Inte

rven

tio

n:

rFS

H(n¼

12

1)

or

HP

-h

FS

H(n¼

12

0)

Sta

rtin

gd

ose

:3

00

IUd

aily

then

adju

sted

10

00

0IU

hC

Gw

hen

�4

foll

icle

so

f�

16

mm

.O

ocy

tere

trie

val

36

haf

ter

hC

G.

Em

bry

os

tran

sfer

red

on

day

3.

Lu

teal

ph

ase

sup

po

rt:

Pro

ges

tero

ne

IM5

0m

gd

aily

To

tal

amo

un

to

fF

SH

adm

inis

tere

d;

Am

ou

nt

of

FS

Hp

ero

ocy

te

Day

so

fst

imu

lati

on

E2

atth

ed

ayo

fh

CG

Nu

mb

ero

fo

ocy

tes

coll

ecte

d

Nu

mb

ero

fem

bry

os

ob

tain

ed

Nu

mb

ero

fem

bry

os

tran

sfer

red

Pre

gn

ancy

rate

Imp

lan

tati

on

rate

Ab

ort

ion

rate

No

An

toin

eJM

etal

.,E

SH

RE

,2

00

7In

terv

enti

on

alra

nd

om

ized

,co

ntr

oll

ed,

op

en-l

abel

,in

ves

tiga-

tor

bli

nd

,m

ult

icen

ter

Wo

men

un

der

go

ing

IVF

/IC

SI

cycl

es(n¼

15

0)

Age

18

to3

9yea

rs;

FS

H4

10

IU/m

L;

BM

I4

30

kg

/m2

Lo

ng

dow

n-r

egu

lati

on

wit

hG

nR

Hag

on

ist

Inte

rven

tio

n:

rFS

H(n¼

76

)o

rH

P-h

FS

H

Nu

mb

ero

fo

ocy

tes

retr

ieved

Cli

nic

alp

reg

nan

cyra

te

OH

SS

rate

Yes

522 S. Gerli et al. Gynecol Endocrinol, 2013; 29(6): 520–529

Gyn

ecol

End

ocri

nol D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y U

nive

rsity

of

Cal

gary

on

10/0

4/13

For

pers

onal

use

onl

y.

stu

dy

Co

mp

ute

r-gen

erat

edal

loca

tio

nco

nce

aled

(sea

led

envel

op

es)

less

than

thre

ep

rio

roocy

tere

trie

val

s(n¼

75

)S

tart

ing

do

se:

22

5IU

dai

lyth

enad

just

ed

Oo

cyte

retr

ieval

34

–3

6h

afte

rh

CG

.E

mb

ryo

str

ansf

erre

do

nd

ay2

–3

afte

ro

ocy

tere

trie

-val

.L

ute

alp

has

esu

pp

ort

:n

ot

stat

ed

Do

seo

fg

on

ado

tro

-p

hin

use

d

Tre

atm

ent

du

rati

on

Ab

ate

Aet

al.,

Gy

nec

ol

En

do

crin

ol,

20

09

Inte

rven

tio

nal

ran

do

miz

ed,

con

tro

lled

,o

pen

-lab

el,

mu

ltic

en-

ter

stu

dy

Co

mp

ute

r-gen

erat

edli

st

Wo

men

un

der

go

ing

IVF

/IC

SI

cycl

es(n¼

40

1)

Age

26

to3

8yea

rs;

less

than

thre

ep

rio

roocy

tere

trie

val

s

FS

H4

10

IU/m

L;

BM

I4

30

kg

/m2

�1

0an

tral

foll

icle

s2

–1

0m

min

size

Lo

ng

dow

n-r

egu

lati

on

wit

hd

aily

Gn

RH

ago

nis

t

Inte

rven

tio

n:

rFS

H(n¼

18

6)

or

HP

-h

FS

H(n¼

21

5)

Sta

rtin

gd

ose

:2

25

IUd

aily

then

adju

sted

Oo

cyte

retr

ieval

34

–3

6h

afte

rh

CG

.E

mb

ryo

str

ansf

erre

do

nd

ay2

–3

afte

ro

ocy

tere

trie

-val

.L

ute

alp

has

esu

pp

ort

:P

roges

tero

ne

IM5

0m

gd

aily

Oo

cyte

san

dem

bry

os

qu

alit

yC

lin

ical

pre

gn

ancy

rate

Nu

mb

ero

fo

ocy

tes

retr

ieved

Do

seo

fg

on

ado

tro

-p

hin

use

d

Tre

atm

ent

du

rati

on

No

Bak

erV

Let

al.,

Fer

til

Ste

ril,

20

09

Inte

rven

tio

nal

ran

do

miz

ed,

con

tro

lled

,o

pen

-lab

el,

inves

tiga-

tor

bli

nd

,m

ult

icen

ter

stu

dy

Co

mp

ute

r-gen

erat

edal

loca

tio

nco

nce

aled

(sea

led

envel

op

es)

Wo

men

un

der

go

ing

IVF

/IC

SI

cycl

es(n¼

15

2)

Age

18

to3

9yea

rs;

BM

I1

8–

30

kg

/m2;

less

than

thre

ep

rio

roocy

tere

trie

val

s;b

asal

FS

H5

10

IU/L

;4

10

antr

alfo

llic

les

Pri

mar

yovar

ian

fail

ure

or

po

or

resp

on

der

s;p

rio

rO

HS

S;

PC

OS

;o

ne

or

bo

thovar

-ie

sin

acce

ssib

lefo

ro

ocy

tere

trie

-val

;ovar

ian

cyst

s4

20

mm

;hy

dro

-sa

lpin

xn

ot

rem

oved

or

ligat

ed;

stag

e3

or

4en

do

met

rio

sis;

oo

cyte

do

nat

ion

Pre

trea

tmen

tw

ith

ora

lco

ntr

acep

tivep

ills

Lo

ng

dow

n-r

egu

lati

on

wit

hd

aily

Gn

RH

ago

nis

t

Inte

rven

tio

n:

rFS

H(n¼

76

)o

rH

P-h

FS

H(n¼

76

)S

tart

ing

do

se:

30

0IU

dai

lyth

enad

just

edu

pto

am

axim

um

do

seo

f4

50

IU/d

hC

Gw

hen�

2fo

lli-

cles

of�

16

mm

.O

ocy

tere

trie

val

34

–3

6h

afte

rh

CG

.E

mb

ryo

str

ansf

erre

d

Nu

mb

ero

fo

ocy

tes

retr

ieved

Pre

gn

ancy

rate

,p

erst

i-m

ula

ted

cycl

e,p

ero

ocy

tere

trie

val

and

per

emb

ryo

tran

sfer

To

tal

do

seo

fF

SH

(IU

s)

stim

ula

tio

nd

ura

tio

n

Nu

mb

ero

ffo

llic

les

41

4m

mo

nth

ed

ayo

fh

CG

inje

ctio

n

Can

cell

atio

nra

te

Fer

tili

zati

on

rate

To

tal

nu

mb

ero

fem

bry

os,

nu

mb

ertr

ansf

erre

d,

fro

zen

and

dis

-ca

rded

Yes

(co

nti

nu

ed)

DOI: 10.3109/09513590.2012.758703 rFSH versus hFSH: meta-analysis 523

Gyn

ecol

End

ocri

nol D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y U

nive

rsity

of

Cal

gary

on

10/0

4/13

For

pers

onal

use

onl

y.

Tab

le1

.C

on

tin

ued

Au

tho

rs,

jou

rnal

,yea

ro

fp

ub

lica

tio

nS

tud

yty

pe

and

des

ign

Par

tici

pan

ts,

incl

usi

on

crit

eria

Excl

usi

on

crit

eria

Inte

rven

tio

ns

Pri

mar

yo

utc

om

esS

eco

nd

ary

ou

tco

mes

Fin

anci

alsu

pp

ort

by

ph

arm

aco

mp

any

and

asse

ssed

on

day

2–

6af

ter

oo

cyte

retr

ieval

Lu

teal

ph

ase

sup

po

rt:

Pro

ges

tero

ne

IM5

0m

gd

aily

Imp

lan

tati

on

rate

Nu

mb

ero

ftr

ans-

ferr

edem

bry

os

Ab

ou

lgh

arM

etal

.,F

erti

lS

teri

l,2

01

0In

terv

enti

on

alp

rosp

ec-

tive

ran

do

miz

edco

n-

tro

lled

stu

dy

All

oca

tio

nco

nce

aled

(sea

led

envel

op

es)

by

an

urs

en

ot

invo

lved

inth

est

ud

y

Wo

men

un

der

go

ing

IVF

/IC

SI

wit

hP

CO

S(R

ott

erd

amcr

iter

ia)

(n¼

84

)

Age

18

to3

7y

rs;

BM

I1

9–

32

kg

/m2;

FS

Hb

elow

10

IU/L

;In

fert

ilit

yd

ue

tom

ale

fact

or,

tub

alfa

cto

r,lo

ng

stan

din

gin

fert

ilit

yn

ot

resp

on

din

gto

oth

erli

nes

of

trea

tmen

t

Pre

vio

us

fail

edIV

F/

ICS

I;g

ener

alm

edic

ald

iso

r-d

ers;

dia

bet

es;

hy

per

ten

sio

n;

thy

roid

dis

tur-

ban

ce;

hy

per

pro

lact

ine-

mia

;in

trau

teri

ne

lesi

on

so

rfi

bro

ids;

end

om

etri

osi

s;n

on

-ob

stru

ctiv

eaz

oo

sper

mia

Lo

ng

dow

n-r

egu

lati

on

wit

hG

nR

Ha

Inte

rven

tio

n:

Met

form

in5

00

mg

twic

ed

aily

Sta

rtin

gd

ose

:1

50

–2

25

IUH

P-h

FS

H(n¼

42

)o

rrF

SH

(n¼

42

),d

epen

din

go

nag

ean

dw

eig

ht

of

pat

ien

t;th

enad

just

ed

ICS

Ian

dem

bry

otr

ansf

er

Nu

mb

eran

dp

erce

nta

ge

of

mat

ure

oo

cyte

sN

um

ber

of

gra

de

Iem

bry

os

Nu

mb

ero

fF

SH

amp

ou

les

Imp

lan

tati

on

rate

OH

SS

rate

Pre

gn

ancy

rate

Yes

Sel

man

HA

etal

.,F

erti

lS

teri

l,2

01

0In

terv

enti

on

alp

rosp

ecti

ve

ran

do

-m

ized

op

en-l

abel

stu

dy

Co

mp

ute

r-gen

erat

edal

loca

tio

nn

ot

stat

ed

Wo

men

un

der

go

ing

firs

tIV

Fcy

cle

(n¼

18

8)

Age

27

to3

8yea

rs;

BM

I2

0–

26

kg

/m2

infe

rtil

ity

du

eto

tub

alfa

cto

r,m

ale

fact

or

or

un

exp

lain

edin

fert

ilit

y

No

tst

ated

Lo

ng

dow

nre

gu

lati

on

wit

hd

aily

Gn

RH

ago

nis

t

Inte

rven

tio

n:

Gro

up

A(n¼

63

)2

25

IUH

P-h

FS

Hfo

rth

efi

rst

6d

star

tin

gfr

om

day

2an

dfo

l-lo

wed

by

22

5IU

rFS

H;

Gro

up

B(n¼

65

)2

25

IUrF

SH

;G

rou

pC

(n¼

60

)2

25

IUH

P-h

FS

H;

FS

Hd

ose

was

adju

sted

wh

enn

eces

sary

Tri

gg

erin

gw

ith

10

00

0IU

hC

GO

ocy

tes

den

ud

edim

med

iate

lyaf

ter

retr

ieval

and

asse

ssed

for

mat

uri

ty

Nu

mb

ero

fm

atu

reo

ocy

tes

Em

bry

oqu

alit

y

Pre

gn

ancy

rate

Imp

lan

tati

on

rate

Do

seo

fF

SH

adm

i-n

iste

red

Tre

atm

ent

du

rati

on

Fer

tili

zati

on

rate

Em

bry

ocl

eavag

era

te

Can

cell

atio

nra

te

Inci

den

ceo

fm

od

er-

ate

or

sever

eO

HS

S

No

Yes

524 S. Gerli et al. Gynecol Endocrinol, 2013; 29(6): 520–529

Gyn

ecol

End

ocri

nol D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y U

nive

rsity

of

Cal

gary

on

10/0

4/13

For

pers

onal

use

onl

y.

Mu

rber

Aet

al.,

Act

aB

iol

Hu

ng

,2

01

1

Inte

rven

tio

nal

ran

do

miz

ed,

con

-tr

oll

ed,

op

en-l

abel

,in

ves

tigat

or

bli

nd

,m

ult

icen

ter

stu

dy

[par

to

fA

nto

ine

etal

.[1

6]

tria

l]

Co

mp

ute

r-gen

erat

edal

loca

tio

nco

nce

aled

(sea

led

envel

op

es)

Wo

men

un

der

go

ing

IVF

/IC

SI

cycl

es(n¼

70

)

Age

18

to3

9yea

rs;

less

than

thre

ep

rio

roocy

tere

trie

val

s

FS

H4

10

IU/m

L;

BM

I4

30

kg

/m2

Lo

ng

dow

n-r

egu

lati

on

wit

hG

nR

Hag

on

ist

Inte

rven

tio

n:

rFS

H(n¼

35

)o

rH

P-h

FS

H(n¼

35

)

Sta

rtin

gd

ose

:2

25

IUd

aily

then

adju

sted

Oo

cyte

retr

ieval

34

–3

6h

afte

rh

CG

.E

mb

ryo

str

ansf

erre

do

nd

ay2

–3

afte

ro

ocy

tere

trie

-val

.L

ute

alp

has

esu

pp

ort

:m

icro

niz

edp

roges

tero

ne

60

0m

gin

trav

agin

ally

Oo

cyte

san

dem

bry

os

qu

alit

yC

lin

ical

pre

gn

ancy

rate

Liv

e-b

irth

rate

Mu

ltip

leges

tati

on

s

Do

seo

fg

on

ado

tro

-p

hin

use

d

Tre

atm

ent

du

rati

on

Ab

bre

via

tio

ns:

IVF¼

invi

tro

fert

iliz

atio

n;

ICS

I¼

intr

a-cy

top

lasm

icsp

erm

inje

ctio

n;

BM

I¼

bo

dy

mas

sin

dex

;H

P-h

FS

H¼

hig

hly

pu

rifi

edh

um

anfo

llic

lest

imu

lati

ng

ho

rmo

ne;

rFS

H¼

reco

mb

inan

tfo

llic

lest

imu

lati

ng

ho

rmo

ne;

IUI¼

intr

aute

rin

ein

sem

inat

ion

;O

HS

S¼

ovar

ian

hy

per

stim

ula

tio

nsy

nd

rom

e;P

CO

S¼

po

lycy

stic

ov

ary

syn

dro

me;

Gn

RH

a¼

go

nad

otr

op

hin

-rel

easi

ng

ho

rmo

ne

ago

nis

t;h

CG¼

hu

man

cho

rio

nic

go

nad

otr

op

hin

.

DOI: 10.3109/09513590.2012.758703 rFSH versus hFSH: meta-analysis 525

Gyn

ecol

End

ocri

nol D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y U

nive

rsity

of

Cal

gary

on

10/0

4/13

For

pers

onal

use

onl

y.

Tree-Age Pro 2006 Suite Software, release 0.1 (TreeAgeSoftware, Inc., Williamstown, MA).

Results

Primary outcomes

Live-birth rate

A forest plot presents the individual and summary OR of thestudies. There was no significant difference in overall live-birthrate OR for rFSH versus HP-hFSH in six trials that reportedthis data [15,17,18,20–22] with a non-significant trend in thefavor of HP-hFSH (pooled OR¼ 0.84 95% CI: 0.63–1.11,p¼ 0.219, random effects). The heterogeneity test was non-significant: Breslow–Day¼ 1.33, p¼ 0.931; Cochran Q¼ 1.33,p¼ 0.931; I2¼ 0% (95% CI¼ 0% to 61%), indicating that therewas no statistical inconsistency in the studies on the finding ofbenefit towards rFSH compared with HP-hFSH. Begg’s andEgger’s tests, used to look for publication bias, indicated thatpublication and related biases were unlikely (Begg–Mazumdar:Kendall’s � , p¼ 0.719; Egger: bias p¼ 0.202; Horbold–Egger:bias p¼ 0.198). The sensitivity analysis performed with Tobias’method excluded the excessive influence of any study(Figure 1).

Clinical pregnancy rate

There was no significant difference in combined clinicalpregnancy rate OR for rFSH versus HP-hFSH, with a non-significant trend in the favor of HP-hFSH. Seven studies reportedthis data [14–18,21,22] (pooled OR¼ 0.85, 95%CI: 0.68–1.1.07,p¼ 0.162, random effects). No heterogeneity was present:Breslow–Day¼ 1.41, p¼ 0.965; Cochran Q¼ 1.41, p¼ 0.965;I2¼ 0% (95% CI¼ 0% to 58.5%). Begg’s and Egger’s tests (Begg–Mazumdar: Kendall’s � p¼ 0.773; Egger: bias p¼ 0.619;Horbold–Egger: bias p¼ 0.610), and sensitivity analysis per-formed with Tobias’ method confirmed that it was unlikely tohave been a source of bias in the present meta-analysis.

Secondary outcomes

No significant differences were observed between rFSH and HP-hFSH treatment in terms of number of oocytes retrieved, numberof mature oocytes and days of stimulation. The random effectsmodel for the calculation of the WMD was used for all treatmentcycle parameters because of significant heterogeneity.

In particular, these are the results regarding the total number ofoocytes in pooled WMD: �0.03 95% CI: �0.67 to 0.602;p¼ 0.921. Seven studies offered data for this outcome measure[14–18,21,22].

Study rFSH HP-hFSH OR (random) Weight (random)x/n x/n 95%CI %

41/134 52/133 0.69 (0.40-1.17) 31.6

Mohamed M et al. 18/129 20/128 0.88 (0.41-1.85) 17.0

29/76 29/76 1.00 (0.49-2.03) 18.9

20/42 19/42 1.10 (0.43-2.83) 11.0

18/65 20/60 0.77 (0.33-1.76) 13.9

Murber A et al., 2011 10/35 11/35 0.87 (0.28-2.74) 7.6

136/481 151/474 0.84 (0.63-1.11) 100

Favours HP-hFSH Favours rFSH

Odds ratio meta-analysis plot [random effects]

Selman HA et al., 2002

Baker VL et al., 2009

Selman H et al., 2010

Aboulghar M et al., 2010

Combined (random)

0.2 0.5 1 2 5

odds ratio (95% confidence interval)

Study rFSH HP-hFSH OR (random) Weight (random)x/n x/n 95%CI %

49/134 61/133 0.68 (0.40-1.14) 21.5

Mohamed Ma et al. 21/129 23/128 0.89 (0.44-1.8) 12.2

21/76 22/75 0.92 (0.45-1.98) 10.3

50/186 60/215 0.95 (0.60-1.51) 26.6

34/76 37/76 0.85 (0.43-1.70) 12.7

22/42 21/42 1.05 (0.41-2.67) 7.2

21/65 23/60 0.77 (0.34-1.71) 9.5

218/708 247/729 0.85 (0.68-1.07) 100

Favours HP-hFSH Favours rFSH

Antoine JM et al., 2007

Abate A et al., 2009

Odds ratio meta-analysis plot [random effects]

Selman HA et al., 2002

Baker VL et al., 2009

Aboulghar M et al., 2010

Selman H et al., 2010

Combined (random)

0.2 0.5 1 2 5

odds ratio (95% confidence interval)

(a)

(b)

Figure 1. OR of live-birth rate (a) and clinical pregnancy rate (b) per randomized woman.

526 S. Gerli et al. Gynecol Endocrinol, 2013; 29(6): 520–529

Gyn

ecol

End

ocri

nol D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y U

nive

rsity

of

Cal

gary

on

10/0

4/13

For

pers

onal

use

onl

y.

The mean number of mature oocytes was not significantlydifferent between patients treated with rFSH versus HP-hFSH(WMD: 0.466, 95% CI: �0.01 to 0.93; p¼ 0.06). This result isbased on the analysis of five studies ([15–17,21,22].

The duration of stimulation was not significantly differentbetween patients treated with rFSH as compared withthose treated with HP-hFSH. Seven trials reported this data[14–18,21,22] (combined WMD: 0.351 days, 95% CI: �0.05 to0.75; p¼ 0.085).

Cost-effectiveness analysis

The estimated costs with each ovarian stimulation strategydemonstrated that HP-hFSH was a more cost-effective strategycompared with rFSH (E7174 versus E2056) with a strongdominance of the ICER. In a sensitivity analysis, when theeffectiveness was varied along different pregnancy rates of rFSH,there was no pregnancy rate at which rFSH became more cost-effective than HP-hFSH and only for a cost of cycle lower thanE657, the recombinant preparation could have been recom-mended as more cost-effective (Figure 2).

Discussion

The clinical efficacy of commercial gonadotropin preparationshas been the subject of a debate during recent years. Each FSHpreparation is characterized by a specific isoform profile. Theisoelectric point (pI) and its effect on different FSH preparationshave been widely investigated: less-acidic isoforms (pI46–8)exhibit high in vitro bioactivity, but they have a shorter circulatinghalf-life than acidic FSH isoforms (pI¼ 4–5) [32,33]. The longer

half-life of the acidic isoforms results in more estrogenicfollicles and follicular development, and maturation and estradiolsecretion [34].

Evident differences between recombinant and human FSHwere recognized: rFSH containing a higher proportion of less-acidic isoforms, whereas human FSH contains a higher proportionof acidic forms [8]. Even though only human (hFSH, HP-hFSH,hMG) or only HP-hFSH (Metrodin HP�, Fostimon�) preparationsare considered, a significant difference of acidic or less-acidicisoforms is present [9].

To explore whether a specific FSH isoform composition affectseffectiveness of a preparation for ovarian stimulation in womenundergoing IVF or ICSI treatment cycles, data of randomizedclinical trials comparing recombinant gonadotropin Gonal-F�

(rFSH) with the high-purified human gonadotropin Fostimon�

(HP-hFSH) were included in the meta-analysis.The clinical results of this study show that there is no proof of

a difference in OR of live–birth rate and clinical pregnancy rate. Ithas to be noted that these results should be viewed with caution,since a clinically significant difference has not been reached.Evaluation of secondary outcomes such as the number of retrievedoocytes and the number of mature oocytes confirms the overlap ofresults between rFSH and HP-hFSH treatment, which are in linewith the lack of difference observed in the probability of livebirth. The number of days of stimulation is also not statisticallydifferent.

All included studies used computer-generated randomizationwith a proper method of allocation concealment. Our meta-analysis has some limitations: (1) the relatively small number ofstudies and (2) the relatively small overall sample size (n¼ 1437),but these results confirm previous reports suggesting that human-derived FSH and rFSH are equivalent in terms of clinicalefficacy. Our data are in agreement with the most recent Cochraneintervention review [6]. Since the preceding outcome analyses donot reveal an advantage foreither rFSH or HP-hFSH products, thecost-effectiveness analysis acquires added importance.

Cost-minimization analyses are sometimes proposed as a morereliable method of economic investigation. Dakin et al. recentlyreported that conducting cost-minimization analysis rather thancost-effectiveness analysis introduces bias into uncertaintyestimates [35]. As the difference in cost between the two groupswas so large, we decided to adopt the cost-effectiveness analysis.

A detailed analysis of different cost components per treatmentcycle demonstrates that hormonal stimulation is a costly step ofIVF/ICSI procedures (in some countries about 60% of the totalcost) [36]. The use of more expensive products, however, does notnecessarily determine increased total treatment costs and cost-ineffectiveness for healthcare expenditure. There is a possibilitythat more expensive medications lead to such an improvement ofefficacy, or live-birth rate, that the total cost per establishedpregnancy may be significantly reduced.

Hence, it is extremely important to assess the cost of eachpregnancy obtained with both treatment options and in conse-quence, it is important to suggest the use of the more cost-effective drug. The cost-effectiveness analysis of the rFSHcompared with hFSH has been the objective of numerous studies,but the efforts undertaken to reveal which preparation was morecost-effective failed to achieve conclusive results [37,38].

Computer-simulated clinical models have been developed toperform pharmacoeconomic studies, creating a decision tree inwhich a certain cycle of treatment is adopted [39–42].

The present study provide the proof that rFSH never representsa cost-effective alternative to HP-hFSH.

The elaboration of the sensitivity analysis of the present studywhere the effectiveness was varied along different pregnancy ratesof rFSH, indicated that rFSH would be more cost-effective than

0

2,000

4,000

6,000

8,000

10,000

12,000

0 0.2 0.4 0.6 0.8 1

CE

effectiveness rFSH

0

1,000

2,000

3,000

4,000

5,000

6,000

7,000

8,000

9,000

0 500 1,000 1,500 2,000 2,500

CE

cost per cycle of rFSH

hFSH

rFSH

hFSH

rFSH

(a)

(b)

Figure 2. Sensitivity analysis on pregnancy rate (a) and cost per cycle (b)of rFSH (costs are expressed in E).

DOI: 10.3109/09513590.2012.758703 rFSH versus hFSH: meta-analysis 527

Gyn

ecol

End

ocri

nol D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y U

nive

rsity

of

Cal

gary

on

10/0

4/13

For

pers

onal

use

onl

y.

HP-hFSH at a threshold of E657 per cycle or at a price of E0.18/IU, with a hypothetical reduction of the price by 67%. This mightbe the acceptable cost at which the rFSH would be more cost-effective than HP-hFSH.

An interesting aspect of sensitivity analysis is the potentialpractical application, relative to each country, to understand whatmay be the cost per cycle for which there may be convenience inthe use of the more expensive drug (rFSH). The relatively highcost of rFSH, compared with human-derived gonadotropins,represents a major problem in countries where the patient has topartially or totally pay for the expense of the drugs [43]. However,even in countries where the cost is fully reimbursed by nationalhealth services, the high social economic request for infertilitydrugs is an increasing problem [44].

In many developed countries, important issues are both to treatinfertility and at the same time to address the rising costs relatedto this treatment: therefore it is important to constantly re-evaluatethe available options in order to determine the most appropriateand cost-effective model for IVF treatment with pharmacoeco-nomic evaluations [45,46].

Nevertheless conclusions of the present analysis are related tothose countries where ratio between costs of HP-hFSH and rFSHis similar to what reported in this study. It must be emphasizedthat in many countries across the world, prices and ratio betweencosts are very different, with a reduction of the economic gapbetween the two formulations.

In conclusion, our meta-analysis clearly demonstrates that, inwomen undergoing ovarian stimulation in IVF or ICSI cycles, theuse of HP-hFSH (Fostimon�) provides the same clinical outcomesas the recombinant formulation (Gonal-F�). In the analyzedeconomical scenario HP-hFSH is more cost-effective than rFSH.These findings could orientate the decision-making process whenchoosing the appropriate gonadotropin for ovarian stimulation.

Declaration of interest

The authors report no declarations of interest.

References

1. Out HJ. Recombinant follicle stimulating hormone: gold standard ornot? Reprod Biomed Online 2005;11:536–9.

2. Al-Inany H, Aboulghar MA, Mansour RT, Serour GI. Ovulationinduction in the new millennium: recombinant follicle-stimulatinghormone versus human menopausal gonadotropin. GynecolEndocrinol 2005;20:161–9.

3. European and Israeli Study Group on Highly Purified Menotropinversus Recombinant Follicle-Stimulating Hormone (EISG). Efficacyand safety of highly purified menotropin versus recombinantfollicle-stimulating hormone in in vitro fertilization/intracytoplasmicsperm injection cycles: a randomized, comparative trial. Fertil Steril2002;78:520–8.

4. Al-Inany H, Aboulghar M, Mansour R, Serour G. Meta-analysis ofrecombinant versus urinary-derived FSH: an update. Hum Reprod2003;18:305–13.

5. Coomarasamy A, Afnan M, Cheema D, et al. Urinary hMG versusrecombinant FSH for controlled ovarian hyperstimulation followingan agonist long down-regulation protocol in IVF or ICSI treatment: asystematic review and meta-analysis. Hum Reprod 2008;23:310–15.

6. van Wely M, Kwan I, Burt AL, et al. Recombinant versusurinary gonadotrophin for ovarian stimulation in assisted reproduc-tive technology cycles. Cochrane Database of Syst Rev2011;CD005354.

7. Wide L, Eriksson K, Sluss PM, Hall JE. Serum half-life of pituitarygonadotropins is decreased by sulfonation and increased bysialylation in women. J Clin Endocrinol Metab 2009;94:958–64.

8. Andersen CY, Westergaard LG, van Wely M. FSH isoformcomposition of commercial gonadotrophin preparations: a neglectedaspect? Reprod Biomed Online 2004;9:231–6.

9. Andersen CY, Ezcurra D. What is the clinical relevance of follicle-stimulating hormone isoforms in fertility treatment? Reprod BiolInsights 2011;4:1–10.

10. Giudice E, Crisci C, Altarocca V, O’Brien M. Characterisation of apartially purified human menopausal gonadotropin preparation.J Clin Res 2001;4:27–34.

11. Anserini P, Costa M, Remorgida V, Venturini PL. A prospective,randomized, controlled clinical study of a new subcutaneous,purified, urinary FSH preparation for controlled ovarian hypersti-mulation in vitro fertilization. Gynecol Endocrinol 2000;14:75–80.

12. Al-Inany HG, Abou-Setta AM. Are all human-derived follicle-stimulating hormone products the same? A systematic review andmeta-analysis using direct and adjusted indirect analyses, todetermine whether Fostimon� is more efficient than Metrodin-HP�. Gynecol Endocrinol 2012;28:94–101.

13. Haynes RB, Wilczynski N, McKibbon KA, et al. Developingoptimal search strategies for detecting clinically sound studies inMEDLINE. J Am Med Inform Assoc 1994;1:447–58.

14. Abate A, Nazzaro A, Salerno A, et al. Efficacy of recombinantversus human derived follicle stimulating hormone on the oocyteand embryo quality in IVF-ICSI cycles: randomised, controlled,multi-centre trial. Gynecol Endocrinol 2009;25:479–84.

15. Aboulghar M, Saber W, Amin Y, et al. Prospective, randomizedstudy comparing highly purified urinary follicle-stimulating hor-mone (FSH) and recombinant FSH for in vitro fertilization/intracytoplasmic sperm injection in patients with polycystic ovarysyndrome. Fertil Steril 2010;94:2332–4.

16. Antoine JM, De Mouzon J, Nicollet B, et al. Effectiveness andtolerability of hFSH compared to rFSH in ICSI: the European study.IBSA Satellite Symposium abstract, ESHRE, Lyon. 2007. Availablefrom: http://www.ibsa.ch/it/eshre_2007_lyon_abstracts-3.pdf [lastaccessed November 2011].

17. Baker VL, Fujimoto VY, Kettel LM, et al. Clinical efficacy of highlypurified urinary FSH versus recombinant FSH in volunteersundergoing controlled ovarian stimulation for in vitro fertilization:a randomized, multicenter, investigator-blind trial. Fertil Steril2009;91:1005–11.

18. Mohamed MA, Sbracia M, Pacchiarotti A, et al. Urinary follicle-stimulating hormone (FSH) is more effective than recombinant FSHin older women in a controlled randomized study. Fertil Steril2006;85:1398–403.

19. Moustafa M, Abdelwahed A, Abosekena I, et al. IVF outcomes witheither highly purified FSH vs recombinant FSH in down-regulatednormogonadotrophic women: a prospective comparative study in adeveloping country and meta-analysis. Open Women’s Health J2009;3:11–15.

20. Murber A, Fancsovits P, Ledo N, et al. Impact of highly purifiedversus recombinant follicle stimulating hormone on oocyte qualityand embryo development in intracytoplasmic sperm injection cycles.Acta Biol Hung 2011;62:255–64.

21. Selman HA, De Santo M, Sterzik K, et al. Effect of highly purifiedurinary follicle stimulating hormone on oocyte and embryo quality.Fertil Steril 2002;78:1061–7.

22. Selman H, Pacchiarotti A, El-Danasouri I. Ovarian stimulationprotocols based on follicle-stimulating hormone glycosylationpattern: impact on oocyte quality and clinical outcome. FertilSteril 2010;94:1782–6.

23. Jadad AR, Moore RA, Carroll D, et al. Assessing the quality ofreports of randomized clinical trials: is blinding necessary? ControlClin Trials 1996;17:1–12.

24. DerSimonian R, Laird N. Meta-analysis in clinical trials. ControlClin Trials 1986;7:177–88.

25. Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuringinconsistency in meta-analyses. BMJ 2003;327:557–60.

26. Begg CB, Mazumdar M. Operating characteristics of a rankcorrelation test for publication bias. Biometrics 1994;50:1088–101.

27. Egger M, Davey Smith G, Schneider M, Minder C. Bias in meta-analysis detected by a simple, graphical test. BMJ 1997;315:629–34.

28. Egger M, Smith GD, Phillips AN. Meta-analysis: principles andprocedures. BMJ 1997;315:1533–7.

29. Tobias A. Assessing the influences of a single study in meta-analysis. Statal Tech Bull 1999;47:15–17.

30. Gerli S, Casini ML, Unfer V, et al. Ovulation induction with urinaryFSH or recombinant FSH in polycystic ovary syndrome patients: aprospective randomized analysis of cost-effectiveness. ReprodBiomed Online 2004;9:494–9.

528 S. Gerli et al. Gynecol Endocrinol, 2013; 29(6): 520–529

Gyn

ecol

End

ocri

nol D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y U

nive

rsity

of

Cal

gary

on

10/0

4/13

For

pers

onal

use

onl

y.

31. Gerli S, Bini V, Di Renzo GC. Cost-effectiveness of recombinantfollicle-stimulating hormone (FSH) versus human FSH in intrauter-ine insemination cycles: a statistical model-derived analysis.Gynecol Endocrinol 2008;24:18–23.

32. D’Antonio MD, Borrelli F, Datola A, et al. Biological characteriza-tion of recombinant human follicle stimulating hormone isoforms.Hum Reprod 1999;14:1160–7.

33. Vitt UA, Kloosterboer HJ, Rose UM, et al. Isoforms of humanrecombinant follicle-stimulating hormone: comparison of effects onmurine follicle development in vitro. Biol Reprod 1998;59:854–61.

34. West CR, Carlson NE, Lee JS, et al. Acidic mix of FSH isoforms arebetter facilitators of ovarian follicular maturation and E2 productionthan the less acidic. Endocrinology 2002;143:107–16.

35. Dakin H, Wordsworth S. Cost-minimisation analysis versus cost-effectiveness analysis, revisited. Health Econ 2013;22:22–34.

36. Bouwmans CA, Lintsen BM, Eijkemans MJ, et al. A detailed costanalysis of in vitro fertilization and intracytoplasmic sperm injectiontreatment. Fertil Steril 2008;89:331–41.

37. Hatoum HT, Keye Jr WR, Marrs RP, et al. A Markov model of thecost-effectiveness of human-derived follicle-stimulating hormone(FSH) versus recombinant FSH using comparative clinical trial data.Fertil Steril 2005;83:804–7.

38. Silverberg K, Daya S, Auray JP, et al. Analysis of the costeffectiveness of recombinant versus urinary follicle-stimulating

hormone in in vitro fertilization/intracytoplasmic sperm injectionprograms in the United States. Fertil Steril 2002;77:107–13.

39. Al-Inany HG, Abou-Setta AM, Aboulghar MA, et al. HMG versusrFSH for ovulation induction in developing countries: a cost-effectiveness analysis based on the results of a recent meta-analysis.Reprod Biomed Online 2006;12:163–9.

40. Daya S, Ledger W, Auray JP, et al. Cost-effectiveness modelling ofrecombinant FSH versus urinary FSH in assisted reproductiontechniques in the UK. Hum Reprod 2001;16:2563–9.

41. Mantovani LG, Belisari A, Szucs TD. Pharmaco-economic aspectsof in-vitro fertilization in Italy. Hum Reprod 1999;14:953–8.

42. Romeu A, Balasch J, Ruiz Balda JA, et al. Cost-effectiveness ofrecombinant versus urinary follicle-stimulating hormone in assistedreproduction techniques in the Spanish public health care system.J Assist Reprod Genet 2003;20:294–300.

43. Gerli S, Bini V, Di Renzo GC. Pharmacoeconomy in ART: theimportance of the gonadotrophin choice. Middle East Fertil Soc J2010;15:64–7.

44. Chambers GM, Sullivan EA, Ishihara O, et al. The economic impactof assisted reproductive technology: a review of selected developedcountries. Fertil Steril 2009;91:2281–94.

45. Collins J. An international survey of the health economics of IVFand ICSI. Hum Reprod 2002;8:265–7.

46. Mladovsky P, Sorenson C. Public financing of IVF: a review ofpolicy rationales. Health Care Anal 2010;18:113–28.

DOI: 10.3109/09513590.2012.758703 rFSH versus hFSH: meta-analysis 529

Gyn

ecol

End

ocri

nol D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y U

nive

rsity

of

Cal

gary

on

10/0

4/13

For

pers

onal

use

onl

y.