Recombinant FSH versus highly purified FSH in intrauterine insemination: systematic review and metaanalysis

Transcription

1 Recombinant FSH versus highly purified FSH in intrauterine insemination: systematic review and metaanalysis Roberto Matorras, M.D., Ph.D., a,b,c Carmen Osuna, M.D., a Antonia Exposito, Ph.D., a Lorena Crisol, B.S., a,b and Jose Ignacio Pijoan, M.D., Ph.D. d a Human Reproduction Unit, Department of Obstetrics and Gynecology, Hospital de Cruces, Baracaldo, Vizcaya; b Department of Medical Specialities, University of the Basque Country, Leioa, Vizcaya; c Valencian Institute of Infertility, Bilbao; and d Clinical Epidemiology Unit, Hospital de Cruces, Baracaldo, Vizcaya, Spain Objective: To compare the pregnancy rates (PRs) in intrauterine insemination (IUI) using recombinant FSH (rec-fsh) or highly purified urinary FSH (HP-FSH). Design: Systematic review and metaanalysis. Setting: University hospital. Patient(s): None. Intervention(s): Electronic and manual searches. Main Outcome Measure(s): PR, per first cycle PR and per woman PR. Result(s): Six randomized trials (713 women, 1,581 cycles) were identified. In three the same doses of rec-fsh and HP-FSH were used ( equal dose group), whereas in the other three the ratio HP-FSH:rec-FSH dose was 1.5. The global metaanalysis showed no differences in PRs. The PR per cycle was similar across the 1.5 ratio group (14.51% vs %; relative risk [RR], 0.970; 95% confidence interval [CI], ). However, the metaanalysis of the equal dose group, showed differences in the PR in favor of rec-fsh (16.36% vs %; RR, 1.394; 95% CI, ). Per woman PR analysis showed similar results (41.44% vs %; RR, 1.273; 95% CI, ). Per first cycle PR analysis showed a similar trend, although the difference did not reach significance (RR, 1.434; 95% CI, ). Conclusion(s): Rec-FSH was associated with higher per cycle PR than HP-FSH, when used at the same dose, whereas the PR were similar when the dose of rec-fsh was 50% lower. (Fertil Steril Ò 2011;95: Ó2011 by American Society for Reproductive Medicine.) Key Words: IUI, insemination, metaanalysis, rec-fsh, HP-FSH, urinary FSH, pregnancy rates A number of methodological differences have been reported concerning intrauterine insemination (IUI) studies (1). Regarding ovarian stimulation, gonadotrophin stimulation results in higher pregnancy rates (PRs) than does no stimulation or clomiphene stimulation (2, 3). Moreover, a number of drugs have been proposed, alone or in combination. Concerning FSH-only protocols, they have received a lot of attention in IVF. In an early metaanalysis, it was concluded that recombinant FSH (rec-fsh) was associated with significantly higher PRs than either purified or highly purified urinary FSH (HP-FSH) in IVF cycles (4). In a posterior metaanalysis (5), although a trend to higher PRs was reported in rec-fsh compared with purified FSH and also with HP-FSH, significance was not reached. This type of metaanalysis has not been performed for IUI. The aim of this study was to conduct a metaanalysis concerning PRs in IUI studies using only FSH stimulation protocols. MATERIAL AND METHODS We performed a systematic review of the literature concerning randomized and prospective trials on IUI in which rec-fsh and urinary FSH were Received October 13, 2010; revised January 27, 2011; accepted February 13, 2011; published online March 23, R.M. has nothing to disclose. C.O. has nothing to disclose. A.E. has nothing to disclose. L.C. has nothing to disclose. J.I.P. has nothing to disclose. Reprint requests: Roberto Matorras, M.D., Ph.D., Hospital de Cruces, Paıs Vasco University, Obstetrics and Gynecology, Maria Diaz de Haro 7, 6 iz Bilbao 48013, Spain ( roberto.matorras@osakidetza.net). compared, regardless of the doses used and their ratio. The search was performed in the databases of the National Library of Medicine s MEDLINE and the Cochrane Library with a variety of keywords, including recombinant FSH, urinary FSH, IUI, and intrauterine insemination for the period from 1966 to November Google Scholar was also searched, and cross references were explored. A number of prospective trial registers were also searched. We also reviewed manually the abstract books of the Annual Meetings of the ESHRE and the American Society for Reproductive Medicine from 1998 to Some authors were contacted to clarify specific aspects of their work. The aim of the analysis was to evaluate the PRs obtained by pooling all the studies retrieved and in the various subgroups corresponding to the different HP-FSH/rec-FSH ratios investigated. The following PR analysis is reported: first cycle PR, per cycle PR, and per woman PR. From our search, nine trials were identified in which the PRs using urinary FSH or rec-fsh for ovarian stimulation in IUI were compared (Fig. 1). Only six of these trials fulfilled the inclusion criteria of being randomized and prospective and comparing PR (Table 1). The review included a total of 1,581 cycles and 713 patients (Table 1). All the studies were single-center studies, all the study groups were kept intact, and no crossover occurred. In two studies, intention-to-treat results regarding per woman PR (considering spontaneous pregnancies) were presented (6, 7). The maximum number of cycles to be performed was predefined in all studies: ranging from six to one. No data were provided concerning live birth rates or ongoing PRs in any study. Only two studies provided information concerning allocation concealment, which was partial in both cases (6, 7). The urinary FSH was HP-FSH in all cases. The main difference among the included studies was the ratio of the dose of HP-FSH:rec-FSH used in the arms of the trial. Three studies used the same dose of HP-FSH and /$36.00 Fertility and Sterility â Vol. 95, No. 6, May doi: /j.fertnstert Copyright ª2011 American Society for Reproductive Medicine, Published by Elsevier Inc.

2 FIGURE 1 Flow diagram of the records. Analysis of rec-fsh vs HP-FSH in IUI. rec-fsh in both arms. Such studies have been designated as equal dose studies. In two of these studies, stimulation was performed with 150 IU FSH (6, 7); in one study, the doses were 75 or 150 UI (8). In the other three reports (9 11) the ratio of the HP-FSH:rec-FSH was 1.5. In one of these studies, 150 IU HP-FSH was compared with 100 of rec-fsh (9) and in the remaining two 75 IU HP-FSH were compared with 50 IU rec-fsh (10, 11). In addition, there were some other differences in cycle management (Table 1) and regarding inclusion criteria. In one study (11), only women with polycystic ovary syndrome were included (Supplemental Text 1 and 2). RESULTS Per Cycle PR In one trial (11), the per cycle PR (PCPR; Fig. 2) was higher in the HP-FSH group, but without statistical significance. In the remaining five trials, PCPRs were higher in the rec-fsh group, although significance was reached only in one of them (8). When all the results were pooled, the PCPRs in rec-fsh population was 16.08% vs % in the HP-FSH group (relative risk [RR], 1.17; 95% confidence interval [CI], ). When the metaanalysis was restricted to the studies using the same FSH dose in rec-fsh and HP-FSH groups (6, 7, 8), the PCPR was higher in the rec-fsh group in all cases, although significance was reached in only one of them (8). When the results of these three studies were pooled, the PCPR was 16.36% in the rec-fsh group, significantly higher than the 12.31% of the HP-FSH group (RR, 1.394; 95% CI, ). When the metaanalysis was performed on just the studies in which the HP-FSH:rec-FSH ratio was 1.5, the PCPR was favored by HP- FSH in one trial (10) and in two by rec- FSH (9, 11), all of them without statistical significance. When these results were pooled, the PCPRs in the rec-fsh and in HP-FSH groups were highly similar (14.51% vs %; RR, 0.970; 95% CI, ; Supplemental Text 3 and Supplemental Fig. 1). Per Woman PR The combined metaanalysis of all six trials showed no differences in the per woman PR (PWPR; Fig. 3; 36.49% in rec-fsh vs % in HP-FSH; RR, 1.156; 95% CI, ). Regarding the equal dose studies, all showed higher PWPR with rec-fsh, although the differences were statistically significant in only one (8). When the results of the equal dose studies were pooled, the PWPR with rec-fsh was 41.44% compared with 31.55% in the HP-FSH group. When the RR was calculated by means of the random effects model, the RR was 1938 Matorras et al. Effectiveness of recombinant and urinary FSH Vol. 95, No. 6, May 2011

3 Fertility and Sterility â 1939 TABLE 1 Included trials. Reference Matorras et al., 2000 (6) Pares et al., 2002 (7) Isaza et al., 2003 (9) Gerli et al., 2004 (10) Gerli et al., 2004 (11) Demirol and Gurgan, 2007 (8) No. of cycles (no. of patients) 155 HP-FSH (46), 139 rec FSH (45) 220 HP-FSH (61), 165 rec FSH (55) 106 HP-FSH (53), 118 rec FSH (55) 67 HP-FSH (32), 71 rec FSH (35) 182 HP-FSH (82), 197 rec FSH (88) 80 HP-FSH (80), 81 rec FSH (81) HP-FSH:rec FSH dose ratio n.s. ¼ no significant differences; ITT ¼ intention to treat; BMI ¼ body mass index. Ovarian management Pregnancy rate/first cycle Pregnancy rate/cycle Pregnancy rate/ woman IU HP-FSH or rec FSH HP-FSH: 11/46 (23.91%) HP-FSH: 24/155 (15.48%) HP- FSH (ITT): 24/46 (52.17%) Two or more follicles R17 mm; E 2 > 400 pg/ml rec FSH: 13/45 (28.88%) rec FSH: 25/139 (17.99%) rec FSH (ITT): 26/45 (57.77%) 5,000 IU hcg; single IUI at 36 h; up to six IUI cycles n.s. n.s. n.s IU HP-FSH or rec FSH HP-FSH: 7/61 HP-FSH: HP- FSH (ITT): (11.47%) 21/220(9.54%) 24/61(39.34%) Al least 1 follicle R18 mm rec FSH: 7/55 rec FSH: 22/165 rec FSH (ITT): (12.73%) (13.33%) 28/55 (50.90%) E 2 >150pg/ml n.s. n.s. n.s. 10,000 IU hcg 2 IUI at 20 and 40 h Up to 6 IUI cycles IU HP-FSH or 100 IU HP-FSH: 11/53 HP-FSH: HP- FSH: rec FSH (20.75%) 22/106 (20.75%) 22/53 (42.30%) At least 1 follicle R18 mm rec FSH: 11/55 rec FSH: rec FSH: (20.00%) 25/118 (21.18%) 25/55 (45.45%) 5,000 IU hcg, two IUI at 12 n.s. n.s. n.s. and 36 h, up to four IUI cycles IU HP-FSH or 50 IU HP-FSH: 16/82 HP-FSH: HP-FSH: rec FSH (19.51%) 8/67 (11.94%) 8/32 (25%) At least one follicle ¼ 18 mm rec FSH: 14/88 rec FSH: rec FSH: (15.91%) 9/71 (12.67%) 9/35 (25.71%) 10,000 IU hcg, single IUI at h, n.s. n.s. n.s. up to three IUI cycles IU HP-FSH or 50 IU HP-FSH: 5/32 HP-FSH: HP-FSH: rec FSH (15.62%) 23/182 (12.63%) 23/82 (28.04%) At least one follicle >18 mm rec FSH: 5/35 rec FSH: rec FSH: (14.29%) 22/197 (11.16%) 22/88 (25.0%) 10,000 IU hcg, single IUI at h, n.s. n.s. n.s. up to three IUI cycles 1 75 IU HP-FSH or rec FSH HP-FSH: 11/80 HP-FSH: HP-FSH: if BMI <25kg/m 2, (13.75%) 11/80 (13.75%) 11/80 (13.75%) 150 IU HP-FSH or rec FSH if BMI R25kg/m 2 One or more follicles R16 mm rec FSH: 21/81 rec FSH: rec FSH: (25.93%) 21/81 (25.93%) 21/81 (25.93%) 10,000 IU hcg, single IUI at 36 h; P<0.05 P<0.05 P<0.05 only one IUI cycle

4 FIGURE 2 Effect on the per cycle pregnancy rate of rec-fsh vs. HP-FSH according the ratio of HP-FSH:rec-FSH dose administered. FIGURE 3 Effect on the per woman pregnancy rate of rec-fsh vs. HP-FSH according the ratio of HP-FSH:rec-FSH dose administered Matorras et al. Effectiveness of recombinant and urinary FSH Vol. 95, No. 6, May 2011

5 1.273 (95% CI, ). When the fixed effects model was applied, the RR was (95% CI, ). Regarding the 1.5 ratio three trials studied, the RR was close to 1 in all of them (ranging ). When the results of these three studies were pooled, their resulting PWPR with rec-fsh was 31.46%, similar to the 31.74% in the HP-FSH group (RR, ; 95% CI, ; Supplemental Text 4, Supplemental Table 1). DISCUSSION There are remarkable differences in IUI methodology (1). Many of these differences are present in the studies reviewed: differences in the inclusion criteria, starting dose of FSH, number of inseminations per cycle, timing of the insemination, timing and criteria for the triggering of ovulation and hcg dose, and the maximum number of cycles performed, among others. Although that heterogeneity could be seen as a difficulty for the metaanalysis, in our opinion it is a true reflection of the real world of IUI patients where a plethora of different protocols are used. In IVF studies, it is generally accepted that the endpoint is the PR or delivery rate for each IVF cycle, although some disagreement exists regarding the consideration of cycles with surplus cryopreserved embryos. In IUI, because PCPRs are considerably lower than in IVF, and because several cycles are performed over a short period of time, there is no agreement concerning the best way to present results. Considering only PWPR would not show the benefit of treatments achieving pregnancy early, requiring fewer IUI cycles to reach pregnancy. Moreover, considering only PCPR has some methodologic problems from a statistical perspective, because not all the patients receive the same number of cycles. Focusing on first cycle PR precludes this problem, but does not provide information regarding cumulative pregnancies. Indeed, as only a proportion of IUI cycles are first cycles, the size of the population available for study decreases. To obtain a more accurate picture of the situation, we have included all three of these endpoints in our analysis: PWPR, PCPR, and per first cycle PR. The different gonadotrophin preparations and combinations have received a lot of attention in IVF cycles, and a number of articles have been published, in many cases with conflicting reports (4, 5) possibly resulting from the different inclusion criteria (12). However, ovarian stimulation protocols have received less attention in IUI. In one recent report it was concluded that HP-hMG was not inferior than rec-fsh regarding clinical PR, while the incidence of OHSS was lower (13). We have found only six randomized trials addressing the topic of the comparison of PR with rec-fsh to HP-FSH in IUI. Of them, the same dose of rec-fsh and HP-FSH was used in three (6 8). Better PCPRs were reported with rec-fsh in all, although statistical significance was reached in only one (8). The metaanalysis was performed involving only studies with the same doses of recombinant and urinary FSH; it showed that PCPRs were significantly higher with rec-fsh. Specifically, from one treatment to the other there was an increase of 2.52% in PR, this being a relative increment of 18.58%. When the metaanalysis was focused on the study of PWPR, the same trend toward higher PRs in rec-fsh was observed, the results being very close to the significance limit (95% CI with the random effects model, , and with the fixed effects model). When the metaanalysis was focused on the PR obtained in the first IUI cycle, a similar trend was observed, although it did not reach statistical significance. Concerning the other studies in which the rec-fsh dose was 50% lower than the urinary FSH dose, the metaanalysis showed that the pooled PRs were highly similar (15.93% and 14.51%), as was the per first cycle PR. Regarding adverse effects, multiple pregnancy, OHSS, and cancellations because of hyperresponsiveness were similar in rec-fsh and HP-FSH. Our metaanalysis has some limitations: [1] the clinical heterogeneity of studies; [2] the fact that three different units of analysis (PR per cycle, per first cycle, and per woman) were studied, all three exhibiting similar trends, but with significant differences only being reached in one; and [3] the relatively small number of studies and total number of cases under analysis. Our data are in agreement with other authors suggesting a higher bioactivity of rec-fsh (14 17). It is not yet fully understood why rec-fsh should be more effective than urinary preparations of FSH, although differences in pharmaceutical formulation and the presence of slightly more basic isoforms of FSH in rec-fsh (18) have been proposed. It has been suggested that another factor could be a difference between follitropin alpha and beta in terms of bioactivity (19). When we performed a metaanalysis restricted to only one kind of rec-fsh, the subgroups obtained coincided with those resulting from the dose analysis. That is, equal dose studies were performed with follitropin alpha, whereas 1.5 ratio studies were performed with follitropin beta. Thus, the first conclusion was that higher per cycle PRs were obtained with follitropin alpha than with the same dose of HP-FSH. Second, with follitropin beta, the same per cycle and per woman PR were obtained as with HP-FSH doses 50% higher. In summary, our metaanalysis revealed that IUI follicular stimulation with rec-fsh achieved higher per cycle PR than when HP- FSH was used at the same dose. When the dose of rec-fsh was 50% lower than the HP-FSH dose, pregnancy rates were similar. However, more studies are required, particularly given the heterogeneity of the trials under analysis, in addition to the fact that a number of the endpoints studied come close to the limits of significance. REFERENCES 1. Matorras R, Corcostegui B, Perez C, et al. Sperm morphology analysis (strict criteria) in male infertility is not a prognostic factor in intrauterine insemination with husband s sperm. Fertil Steril 1995;63: Guzick DS, Carson SA, Coutifaris C, et al. Efficacy of superovulation and intrauterine insemination in the treatment of infertility. National Cooperative Reproductive Medicine Network. N Engl J Med 1999;340: Matorras R, Diaz T, Corcostegui B, et al. Ovarian stimulation in intrauterine insemination with donor sperm: a randomized study comparing clomiphene citrate in fixed protocol versus highly purified urinary FSH. Hum Reprod 2002;17: Daya S, Gunby J. Recombinant versus urinary follicle stimulating hormone for ovarian stimulation in assisted reproduction. Hum Reprod 1999;14: Al-Inany H, Aboulghar M, Mansour R, Serour G. Meta-analysis of recombinant versus urinaryderived FSH: an update. Hum Reprod 2003;18: Matorras R, Recio V, Corcostegui B, Rodrıguez- Escudero FJ. Recombinant human FSH versus highly purified urinary FSH: a randomized study in intrauterine insemination with husbands spermatozoa. Hum Reprod 2000;15: Pares P, Bordas JR, Sak MJ, et al. Recombinant FSH versus urinary FSH in ovarian stimulation for intrauterine insemination with husband s sperm. Prospective and randomised study. [Spanish]. Rev Iberoam Fertil 2002;19: Demirol A, Gurgan T. Comparison of different gonadotrophin preparations in intrauterine insemination cycles for the treatment of unexplained infertility: a prospective, randomized study. Hum Reprod 2007;22: Isaza V, Requena A, Garcıa-Velasco JA, et al. Recombinant vs. urinary follicle-stimulating hormone in couples undergoing intrauterine insemination. A randomized study. J Reprod Med 2003;48: Gerli S, Casini ML, Unfer V, et al. Ovulation induction with urinary FSH or recombinant FSH in poly- Fertility and Sterility â 1941

6 cystic ovary syndrome patients: a prospective randomized analysis of cost-effectiveness. Reprod Biomed Online 2004;9: Gerli S, Casini ML, Unfer V, et al. Recombinant versus urinary follicle-stimulating hormone in intrauterine insemination cycles: a prospective, randomized analysis of cost effectiveness. Fertil Steril 2004;82: Matorras R, Prieto B, Exposito A, et al. Supplementation with mid-follicular LH in women aged over 35 years undergoing COS in cycles of ICSI: a randomized controlled study. Reprod Biomed Online 2009;19: Sagnella F, Moro F, Lanzone A, et al. A prospective randomized noninferiority study comparing recombinant FSH and highly purified menotropin in intrauterine insemination cycles in couples with unexplained infertility and/or mild-moderate male factor. Fertil Steril 2011;95: Recombinant Human FSH Study Group. Clinical assessment of recombinant human follicle-stimulating hormone in stimulating ovarian follicular development before in vitro fertilization. Fertl Steril 1995;63: Out HJ, Mannaerts BMJL, Driessen SGAJ, Bennink HJ. A prospective, randomized, assessor-blind, multicenter study comparing recombinant and urinary folliclestimulating hormone (Puregon vs Metrodin) in in vitro fertilization. Hum Reprod 1995;10: Balasch J, Fabregues F, Pe~narrubia J, et al. Follicular development and hormonal levels following highly purified or recombinant follicle-stimulating hormone administration in ovulatory women and WHO group II anovulatory infertile patients. J Assist Reprod Genet 1998;15: Howles CM, Loumaye E, Giroud D, Luyet G. Multiple follicular development and ovarian steroidogenesis following subcutaneous administration of a highly purified urinary FSH preparation in pituitary desensitized women undergoing IVF: a multicenter European phase III study. Hum Reprod 1994;9: Andersen CY, Westergaard LG, van Wely M. FSH isoform composition of commercial gonadotrophin preparations: a neglected aspect? Reprod Biomed Online 2004;9: Orvieto R, Nahum R, Rabinson J, et al. Follitropin-alpha (Gonal-F) versus follitropin-beta (Puregon) in controlled ovarian hyperstimulation for in vitro fertilization: is there any difference? Fertil Steril 2009;91: WHO laboratory manual for the examination of human semen and sperm-cervical mucus interaction, Cambridge University Press, Matorras et al. Effectiveness of recombinant and urinary FSH Vol. 95, No. 6, May 2011

7 SUPPLEMENTAL TEXT 1 One study (6) used 150 IU HP-FSH (Metrodin HP; Serono, Geneva, Switzerland) or rec-fsh (Gonal f; Serono), and ovulation was triggered with 5,000 IU hcg when at least two or more follicles measured R17 mm and E 2 levels >400 pg/ml were obtained. One single insemination was performed 36 hours after hcg, and up to six cycles were performed. In the second of these trials (7), 150 IU HP-FSH (Neo-Fertinorm; Serono; in most countries marketed under the name of Metrodin HP) or rec-fsh (Gonal f) were administered. When at least one follicle measured R18 mm and E 2 >150 pg/ml were obtained, 10,000 IU hcg was administered. Two inseminations were performed 20 and 40 hours after hcg, and treatment was continued for up to six cycles. In the third study (8), ovarian stimulation was performed with rec- FSH (Gonal f) or HP-FSH (Metrodin-HP): 75 UI FSH if body mass index (BMI) was <25 kg/m 2 and 150 UI if BMI was R25 kg/m 2 (10). Another arm of the treatment consisted of hmg stimulation (not considered in the present metaanalysis). When one or more follicles R16mm were obtained, ovulation was triggered with 10,000 IU hcg. A single insemination was performed 36 hours after hcg administration. In this study, only one cycle was performed per patient. The other three trials used different doses of HP-FSH than rec- FSH for ovarian stimulation (9 11) (Table 1). In one of these trials (9), the doses compared were 150 IU HP-FSH (Neofertinorm; Serono; marketed in the majority of countries as Metrodin HP) and 100 IU rec FSH (Puregon; Organon, Oss, the Netherlands). When at least one follicle R18mm was obtained, 5,000 IU hcg were administered. Two inseminations were performed 12 and 36 hours after hcg. Up to four IUI cycles were performed. The two additional trials were published the same year by the same group. In the first trial (10), the ovarian stimulation was performed with 75 IU HP-FSH (Fostimon; AMSA, Rome, Italy) or 50 IU rec-fsh (Puregon). When one follicle R18mm was obtained, ovulation was triggered with 10,000 IU hcg and a single insemination was performed hours after hcg. Up to three IUI cycles were performed. The second was specifically focused on women with polycystic ovary syndrome (11), with the same stimulation protocol with 75 IU HP-FSH (Fostimon, AMSA) or 50 IU rec-fsh (Puregon; Organon), with 10,000 IU hcg being administered if at least one follicle R18mm was obtained. As before, there was a single insemination hours after hcg administration, and up to three IUI cycles were performed. Regarding inclusion criteria, the first trial (6) included male factor infertility defined as a subnormal sperm analysis according to World Health Organization criteria (20), idiopathic infertility, and failure to obtain pregnancy after six cycles of programmed intercourse. The following pathologies were detected in the women: endometriosis, ovulatory disorders, uterine myomata, hyperprolactinemia, and mild tubal, cervical, or immunologic factors. In the second trial (7), male factor according to WHO criteria and cases of endometriosis, ovulatory disorders, cervical and immunologic factors, and unexplained infertility were included. The third study (9) included four groups of diagnosis: ovulatory dysfunction, endometriosis, male factor, and unexplained infertility. One study included only women with polycystic ovary syndrome (11), whereas inclusion criteria were endometriosis (stage I or II), ovulatory factor, mild male factor, and unexplained infertility in the other study (10). Finally, the last trial (8) included only patients with unexplained infertility. SUPPLEMENTAL TEXT 2 Statistical Analysis Relative risk (RR) was used as the effect estimate of choice. The statistical heterogeneity of results was tested by the chi-square test and the I2 statistic. Results of fixed and random effects models were compared and random-effect estimates were used, producing highly similar results. In general, only data concerning the random effects model will be presented, because it corresponds to the more conservative analysis. However, when the small differences in results obtained with the two models implied a change in the statistical significance, data corresponding to the fixed effect model will also be reported. Funnel plots were used to visually test for potential problems with publication or study quality bias. The Begg adjusted rank correlation test and the Egger regression asymmetry test statistically assessed the potential for such bias. Stata 9.2 for Windows was used to perform these statistical analyses. SUPPLEMENTAL TEXT 3 Per First Cycle PR Concerning the PR obtained in the first cycle, among the equal dose studies, in one of them the PR was significantly higher in the rec FSH group (8), whereas the PRs were somewhat higher in the remaining two (6, 7), but not to a statistically significant degree. On the other hand, the first cycle PR was slightly lower in the rec FSH in all 1.5:1 ratio studies (9 11), but again the differences were not significant (Supplemental Fig. 1). When all the results were pooled using the random effects model, no differences were found between the rec-fsh and the HP-FSH. Considering only the women receiving the same FSH dose, the RR was (95% CI, ; P¼0.10), whereas the RR was when the studies analyzing the 1.5:1 ratio were considered (95% CI, ; Supplemental Fig. 1). SUPPLEMENTAL TEXT 4 Stratification Considering Clinical Parameters There was one trial specifically focused on women with polycystic ovarian disease (11) and another including only idiopathic infertilities (8). The remaining four trials analyzed mixed causes, not reporting separately the PR with the different causes of infertility. The small number of studies precluded a valuable metaanalysis. Regarding the stratification according to the different commercially available gonadotrophins, in three studies the comparison was made between follitropin alpha (Gonal f; Serono) and highly purified urinary FSH (Metrodin; Serono). All of them corresponded to equal dose studies. The resulting metaanalysis is thus exactly the same as that obtained in the equal dose metaanalysis. Moreover, the studies analyzing the 1.5 ratio were all performed with follitropin beta (Puregon), the HP urinary gonadotrophin being Neofertinorm (marketed in many countries as Metrodin HP) in one case (9) and Fostimon (AMSA) in two (10, 11). The results are the same as those obtained in the 1.5-ratio metaanalysis. Side Effects Multiple pregnancy analysis showed similar rates in rec-fsh and HP-FSH both in subtotal analyses (ratio 1:1 and ratio 1.5), as well as in the combined analysis (Supplemental Table 1). The frequency of ovarian hyperstimulation syndrome was similar in rec-fsh and HP-FSH in subtotal analysis (1/470 vs. 0/434 cycles when the ratio was 1:1, and 2/388 vs. 2/355 cycles when the ratio was 1:1.5). Cancellation rate was reported in five trials (6, 7, 9 11). In the equal dose group, the frequency was 2.05% (8/389) in rec-fsh and 2.54% (9/354) in the HP-FSH group. In the 1:1.5 ratio group the cancellation rate because of hyperresponsiveness was 3.87% (15/388 in the rec-fsh group) and 3.67% (13/355) in the HP-FSH group. There were no differences in the cancellation rate because of poor response. Fertility and Sterility â 1942.e1

8 SUPPLEMENTAL FIGURE 1 Effect on the per first cycle pregnancy rate of rec-fsh vs. HP-FSH according the ratio of HP-FSH:rec-FSH dose administered e2 Matorras et al. Effectiveness of recombinant and urinary FSH Vol. 95, No. 6, May 2011

9 SUPPLEMENTAL TABLE 1 Multiple pregnancy rate in IUI: rec-fsh vs. HP-FSH. Reference Rec FSH HP-FSH RR (95% CI) Matorras et al., 2000 (6) 15.4 (4/26) 29.2 (7/24) Pares et al., 2002 (7) 18 (4/22) 14.2 (3/21) Demirol and Gurgan, 2007 (8) 9.5 (2/21) 0 (0/10) Subtotal (10/69) (10/55) 0.90 ( ) Isaza et al., 2003 (9) 21.7 (5/23) 26.3 (5/19) Gerli et al., 2004 (10) 0 (0/9) 0 (0/8) Gerli et al., 2004 (11) 15 (3/20) 13 (3/23) Subtotal (8/52) 16 (8/50) 0.98 ( ) Total (18/121) (18/105) 0.93 ( ) Note: No significant differences. Fertility and Sterility â 1942.e3