RBMOnline - Vol 15 No Reproductive BioMedicine Online; on web 25 September 2007

Transcription

1 RBMOnline - Vol 15 No Reproductive BioMedicine Online; on web 25 September 2007 Many randomized trials have evaluated the use of various pituitary suppression s to improve outcome of poor responders undergoing IVF treatment. A systematic review was conducted of the trials of gonadotrophin-releasing hormone () long,, ant, as well as other pituitary suppression s in poor responders undergoing IVF/intracytoplasmic sperm injection (ICSI) treatment. The search included MEDLINE, EMBASE, Cochrane Library, National Research Register and ISI proceedings, and all randomized controlled trials comparing the various pituitary suppression s in poor responders were included. Study selection, quality appraisal and data extraction were performed independently and in duplicate. The main outcome measures were number of oocytes retrieved, cycles cancelled before oocyte retrieval and pregnancy rates. A total of 680 women considered as poor responders undergoing IVF/ICSI treatment were included in nine randomized controlled trials. The quality of these studies was variable: for example, only three of the studies had clear evidence of allocation concealment. Meta-analyses of the results of the studies did not show a consistent benefit for any one pituitary suppression over the other s in improving outcome measures. Currently available evidence does not favour any one pituitary suppression for women with poor ovarian response undergoing IVF/ICSI treatment. Keywords: long,, ant, meta-analysis, randomized trials, systematic review Poor ovarian response, defined as failure of the development of sufficient number of mature follicles to proceed to oocyte retrieval or yielding only a few oocytes following gonadotrophin stimulation in women undergoing IVF treatment, occurs in 5 25% of women (Keay et al., 1997). Poor ovarian response is likely to be an increasing problem with women delaying childbearing and presenting for treatment later in their reproductive life. In comparison with normal responders, these patients have impaired fertilization rates, lower embryo quality and decreased pregnancy rates (Mahutte and Arici, 2002). Poor responders often have their treatment cycle cancelled because of expected poor outcome (Lashen and Ledger, 1999). This causes emotional distress for the couple, as well as increasing the financial burden on the couple or the service provider. Treatment of poor responders who are undergoing IVF treatment remains a challenge. Various treatment s and interventions have been proposed in an effort to improve ovarian response and IVF outcome in this group of patients. These include different s for pituitary suppression and ovarian stimulation, as well as adjuvant therapies (Shanbhag et al., 2007). Currently, the choice of pituitary suppression s proposed for the management of poor responders is based mainly on the preferences of individual centres or clinicians, 2007 Published by Reproductive Healthcare Ltd, Duck End Farm, Dry Drayton, Cambridge CB3 8DB, UK

2 Article - Pituitary suppression s in poor responders - SK Sunkara et al. with no single protocol considered to be superior over others. The reasons for this variation in practice are likely to include paucity of published evidence and heterogeneity in the trials with regards to issues such as the definition for poor response and clinical protocols used in individual studies. In an attempt to summarize the existing evidence on the various s for pituitary suppression, a systematic review was conducted. MEDLINE (1966-January 2007), EMBASE (1974-January 2007), Cochrane Library and the National Research Register were searched for relevant studies. The search also included ISI Conference Proceedings, which cover over 4.1 million abstracts from over 60,000 conferences since 1990, as well as databases for registration of ongoing and archived randomized controlled trials (RCT) namely ISRCTN register and metaregister for RCT (mrct). A combination of Medical Subject Headings (MeSH) and text words were used to generate two subsets of citations, one including studies of poor ovarian response ( poor response, low response, non response, weak response, failed response ) and the other, studies of IVF and intracytoplasmic sperm injection ( in-vitro fertilization, fertilization in vitro, intracytoplasmic sperm injection, sperm injections intracytoplasmic, reproductive techniques assisted, embryo transfer and embryo implantation ). These subsets were combined using AND to generate a subset of citations relevant to the research question. The reference lists of all known primary and review articles were examined to identify cited articles not captured by electronic searches. Articles frequently cited were used in the Science Citation Index to identify additional citations. Enquiries were also made about unpublished studies from researchers investigating in this field. No language restrictions were placed in any of the searches. Studies were selected if the target population were women with a history of poor ovarian response undergoing IVF or ICSI treatment and the therapeutic interventions were the gonadotrophin-releasing hormone () long (A), the (B), the ant (C), or other pituitary suppression s. Studies were included if they were of randomized design. The primary outcomes were number of oocytes retrieved and cycles cancelled and the secondary outcomes were clinical and ongoing pregnancies. Data were also obtained on duration of gonadotrophin use as a surrogate outcome. Studies were selected in a two-stage process. Firstly, two reviewers scrutinized the titles and abstracts from the electronic searches independently (SKS and JT), and full manuscripts of all citations that definitely or possibly met the predefined selection criteria were obtained. Secondly, final inclusion or exclusion decisions were made on examination of the full manuscripts. In cases of duplicate publication, the most recent and complete versions were selected. The assessment of manuscripts was performed independently by two reviewers (SKS and JT) and any disagreements about inclusion were resolved by consensus or arbitration by a third reviewer (MK). The selected studies were assessed for methodological quality by using the components of study design that are related to internal validity (CRD Report Number 4, 2001). Information on the adequacy of randomization, allocation concealment, blinding, intention to treat analysis and followup rates was sought by examining the full text articles and by contacting the authors if necessary. Study characteristics such as population features and interventions were extracted from each study. Ovarian stimulation is performed in order to increase the number of available oocytes for assisted reproductive technologies. Premature surges of LH can disrupt oocyte maturation, thereby affecting oocyte quality, fertilization rates, endometrial receptivity and pregnancy rates (Griesinger et al., 2004). Pituitary suppression prevents these LH surges and can be achieved either by using s or ants. long With the long, pituitary desensitization with the is commenced in the mid-luteal phase of the previous cycle. After confirmation of ovarian quiescence, gonadotrophins for ovarian stimulation are commenced and continued with a reduced dose of until the administration of human chorionic gonadotrophin (HCG). In the, the is commenced in the early follicular phase of the cycle (day 2 or day 3) followed by gonadotrophin (which is usually commenced a day later). Both the and the gonadotrophin are continued until the administration of HCG. ant The ants competitively block the pituitary receptors and thereby cause immediate suppression of LH (Klingmuller et al., 1993). In the ant, ovarian stimulation with gonadotrophins is commenced in the early follicular phase. The ant is commenced either on day 5 or 6 of stimulation or when the leading follicle is 14 mm. Both the gonadotrophins and the ant are continued until the day of HCG. Long stop With this pituitary desensitization with the is commenced in the mid-luteal phase of the previous cycle. The administration is discontinued on the commencement of gonadotrophin administration. The stop is thought to improve ovarian responsiveness based on a hypothetical effect on the ovary, via receptor or indirectly the vascular network within the ovary (Garcia-Velasco et al., 2000).

3 Article - Pituitary suppression s in poor responders - SK Sunkara et al. From each study, outcome data were extracted in 2 2 tables (for discrete variables) or as means or medians with a measure of variance (such as standard deviation) for continuous variables. For discrete variables, relative risks were pooled using fixed and random effects models, if significant heterogeneity could be excluded. Weighted mean difference (WMD) was calculated for continuous variables using means and standard deviations (SD) from individual studies. As the long, the and the ant are the commonly used s, these formed the focus of the analysis. Heterogeneity of treatment effects was evaluated graphically using forest plot and statistically using chi-squared test. Exploration of clinical heterogeneity was planned using variation in features of the population, intervention, outcome and study quality. All statistical analyses were performed using RevMan 4.2 (The Cochrane Collaboration). Figure 1. Study selection process for the systematic review of pituitary suppression s in patients with poor ovarian response undergoing IVF/ICSI treatment.

4 Article - Pituitary suppression s in poor responders - SK Sunkara et al. The process of literature identification and selection is summarized in Figure 1. Of the 1154 citations identified, 92 were selected during the initial screening, and on examination of manuscripts, nine articles that included a total of 680 women were identified that satisfied the selection criteria for the review. The literature search found one randomized controlled trial comparing the long (A) with the (B) (Weissman et al., 2003), two randomized controlled trials comparing the long (A) with the ant (C) (Cheung et al., 2005; Marci et al., 2005) and four randomized controlled trials comparing the (B) with the ant (C) (Akman et al., 2001; Malmusi et al., 2005; Schmidt et al., 2005; De Placido et al., 2006). It also identified one randomized controlled trial comparing the long with the long stop (Garcia-Velasco et al., 2000), and one randomized controlled trial comparing the ant with IVF cycles not having pituitary suppression with either the or ant (Akman et al., 2000). Two randomized trials were excluded (Dirnfeld et al., 1999; Morgia et al., 2004), as a proportion of participants in these trials contributed more than one cycle to research, making the observations not completely independent from each other. The quality of the nine included trials was generally poor (for example, there was no evidence of allocation concealment in six of the nine studies). These trials were also generally small, leading to imprecision in their findings, even when combined in meta-analyses. Furthermore, there was substantial clinical and statistical heterogeneity amongst the trials, weakening any inferences that could be drawn from these studies. The quality and the main characteristics of the included studies are presented in Tables 1 and 2. Results for the number of oocytes retrieved are summarized in Figure 2. In the trial comparing the long (A) with the (B), a significantly greater number of oocytes were retrieved with the long [WMD 1.35, 95% confidence interval (CI) ; Figure 2(i)]. Meta-analysis of the four randomized controlled trials comparing the (B) with the ant (C) showed a significant benefit with the [WMD 0.48, 95% CI ; Figure 2(ii)]. However, meta-analysis of the two trials comparing the long (A) with the ant (C) showed significantly fewer oocytes retrieved with the long [WMD 1.07, 95% CI 1.92 to 0.21; Figure 2(iii)]. Findings for other comparisons for the outcome of oocytes retrieved are summarized in Table 3. Results for the cancellation rate are summarized in Figure 3 and Table 3. No statistically significant differences were identified in any of the five comparisons. In the trial comparing the long (A) with the (B) there was a higher clinical pregnancy rate with the long, although this effect was not statistically significant (RR 6.55, 95% CI ). A meta-analysis of the trials comparing the (B) with the ant (C) showed no significant difference with clinical pregnancy rates (RR 0.96, 95% CI ) and ongoing pregnancy rates (RR 0.83, 95% CI ). In the trials comparing the long (A) with the ant (C), there was a trend of benefit with the ant for clinical pregnancy rates (RR 2.04, 95% CI ) and ongoing pregnancy rates (RR % CI ), but the results were not statistically significant. Findings for other comparisons for the outcomes of clinical and ongoing pregnancy rates are summarized in Table 3. Table 1. Quality of studies included in the review of pituitary suppression s in poor responders. Study Method of randomization Method of allocation concealment Intention to treat analysis Follow-up rate Clinical pregnancy Ongoing pregnancy Akman et al., 2000 Consecutive number method Unreported Yes >95% >95% Garcia-Velasco et al., 2000 Computer-generated list Opaque sealed envelopes Yes >95% No data Akman et al., 2001 Unclear Unreported Yes >95% >95% Weissman et al., 2003 Computer-generated list Unreported Yes >95% No data Cheung et al., 2005 Computer-generated list Opaque envelopes No >95% >95% Schmidt et al., 2005 Computer-generated list Sealed envelopes No >95% >95% Malmusi et al., 2005 Unclear Unreported No >95% No data Marci et al., 2005 Randomization list Unreported Yes >95% >95% De Placido et al., 2006 Computer-generated list Unreported No >95% >95%

5 Article - Pituitary suppression s in poor responders - SK Sunkara et al. Table 2. Characteristics of the studies included in the review of pituitary suppression s in IVF treatment. Study Participants Intervention Groups Outcomes Akman et al., 2000 n = 40; women who had two failed IVF cycles due to one of the following: FSH >15 miu/ml, serum E2 concentration <500 pg/ml on day of HCG or <4 mature oocytes retrieved n = 20; no pituitary suppression with the or ant n = 36; long n = 24; n = 20; ant Cancelled cycles; number of mature oocytes retrieved; number of embryos transferred; clinical pregnancies; ongoing pregnancies Garcia-Velasco et al., 2000 n = 70; women who had at least one previous IVF cycle cancelled due to 3 follicles 18 mm n = 34; stop n = 24; ant retrieved; number of embryos transferred; clinical pregnancies Akman et al., 2001 n = 48; women who had two failed IVF cycles due to one of the following: FSH >15 miu/ml, serum E2 concentration <500 pg/ml on day of HCG or <4 mature oocytes retrieved retrieved; number of mature embryos transferred; clinical pregnancies; ongoing pregnancies Weissman et al., 2003 n = 60; women with poor response in a previous IVF cycle. Poor response defined as: <5 oocytes, 3 follicles 16 mm on day of cycle cancellation or serum E2 <500 pg/ml on day of HCG. Prior stimulation s involved 300 IU of gonadotrophins daily n = 66; women with poor response in a previous IVF cycle: <3 mature follicles or poor responders defined as women with repeated high basal concentrations of FSH >10 IU/l n = 29; modified n = 33; long n = 31; modified long n = 33; ant retrieved; number of embryos transferred; clinical pregnancies Cheung et al., 2005 Cancelled cycles; duration of stimulation; number of oocytes retrieved; number of mature oocytes retrieved; number of embryos transferred; clinical pregnancies; ongoing pregnancies; live births Schmidt et al., 2005 n = 48; women with poor response in previous IVF cycles. Poor response defined as: peak serum E2 concentration 850 pg/ml or 4 follicles 15 mm on day of HCG. Prior stimulation s involved 300 IU of gonadotrophins daily n = 55 (60 initially recruited); women with poor response in a previous IVF cycle. Poor response defined as: no ovarian response when 300 IU/day of FSH administered for 15 days or <5 oocytes retrieved n = 24; microdose n = 30; n = 24; ant retrieved; number of mature embryos transferred; clinical pregnancies retrieved; number of mature embryos transferred; clinical pregnancies Cancelled cycles; number of embryos transferred; clinical pregnancies; ongoing pregnancies Malmusi et al., 2005 n = 25; ant Marci et al., 2005 n = 60; women with poor response in a previous IVF cycle. Poor response defined as: serum E2 concentration <600 pg/ml on day of HCG and <3 oocytes retrieved after a previous standard long protocol using recombinant gonadotrophin dose of 225 IU/day for stimulation n = 133 (140 initially recruited); women at risk of poor ovarian response when undergoing their first ICSI treatment based on age 37 years or day 2 FSH 9 IU/l n = 30; long n = 30; ant De Placido et al., 2006 n = 67; n = 66; ant Cancelled cycles; number of mature oocytes retrieved; number of embryos transferred; clinical pregnancies; ongoing pregnancies E2 = oestradiol; HCG = human chorionic gonadotrophin.

6 Article - Pituitary suppression s in poor responders - SK Sunkara et al. Table 3. Summary of outcomes for trials comparing various pituitary suppression s. Comparison Outcome Oocytes retrieved (WMD; 95% CI) Cancelled cycles (RR; 95% CI) Clinical pregnancies (RR; 95% CI) Ongoing pregnancies (RR; 95% CI) Duration of stimulation (WMD; 95% CI) Agonist long versus Agonist versus ant Agonist long versus ant Agonist stop versus long Ant versus no pituitary suppression 1.35; ; ; ; ; 95% ; 1.92 to ; ; ; ; ; ; ; ; 4.73 to ; ; ; ; 1.41 to ; ; ; ; Data retrieved for mature oocytes only. CI = confidence interval; RR = relative risk; WMD = weighted mean difference. Figure 2. Oocytes retrieved. (i) Summary of the results for the long (A) versus the (B). (ii) Summary of the results for the (B) versus the ant (C). (iii) Summary of the results for the long (A) versus the ant (C). There was no significant difference in the duration of stimulation between the long (A) versus the (B) (WMD 1.20, 95% CI ) and the (B) versus the ant (C) (WMD 0.11, 95% CI ). Meta-analysis of the trials comparing the long (A) with the ant (C), however, showed that the duration of gonadotrophin stimulation was significantly er with the ant (WMD 4.26, 95% CI 4.73 to 3.78). Findings for other comparisons for the outcome of duration of stimulation are summarized in Table 3. The long is generally considered the standard protocol in most IVF centres (Daya, 1997). With this, pituitary suppression (prior to ovarian stimulation) induced by the suppresses the endogenous LH surge and premature luteinization of oocytes (Hugues and Cedrin-Durnerin, 1998), thus increasing pregnancy rates per IVF treatment cycle.

7 Article - Pituitary suppression s in poor responders - SK Sunkara et al. Figure 3. Cancelled cycles. (i) Summary of the results for the long (A) versus the (B). (ii) Summary of the results for the (B) versus the ant (C). (iii) Summary of the results for the long (A) versus the ant (C). The was introduced in an attempt to improve cycle outcome in poor responders (Garcia et al., 1990; Faber et al., 1998) by avoiding excessive pituitary suppression with the long, while taking advantage of the additional gonadotrophin stimulation provided by the initial flare effect of the. However, some investigators failed to reach similar conclusions (Crammer et al., 1999) and postulated that any potential benefit from the flare effect could be offset by the -induced LH rise in the early follicular phase. The ant avoids profound suppression of the endogenous FSH and LH concentrations in the early follicular phase at the stage of follicular recruitment (Kenigsberg et al., 1984). It was postulated the combination of exogenous and endogenous FSH enables recruitment of more follicles in poor responders. In this review, it was demonstrated that there is insufficient evidence to suggest any one pituitary suppression is associated with improved cycle outcomes, including number of oocytes retrieved, cycle cancellation rate, and clinical and ongoing pregnancy rates. The validity of this finding depends on the methodological rigour of the review and the component primary studies. A prospective protocol was used, and a concerted effort made to find all the evidence. Two independent reviewers assessed study quality and extracted data and the agreement between the two reviewers was high. One major problem usually encountered when synthesizing the evidence in meta-analyses is clinical and methodological heterogeneity between the studies, namely the differences between the studies in the populations they evaluated (definition of poor ovarian response), the dosages of the and ant used, times of commencement and duration of treatment, differences in the way outcomes were defined and reported, as well as important differences in the quality features between the studies. Although these factors were encountered during the analysis, a rigorous exploration of their effect on the study conclusion using techniques such as meta-regression was not feasible due to the limited number of studies identified. In addition, the wide confidence intervals around the relative risks for the outcome measures suggest imprecise results, which could be directly attributed to the relatively small sample sizes of the studies in this review. Furthermore, the results show inconsistency in trial outcomes. For example, the long (A) was shown to be superior to the (B), whilst the (B) was superior to the ant (C). Even though it would be logical to predict the long (A) would be superior to the ant (C), the opposite result was found as the ant (C) yielded a significantly greater number of oocytes and higher clinical and ongoing pregnancy rates in comparison to the long (A) (Figure 2). This inconsistency limits the reliability of existing evidence in informing clinical decision-making while treating poor responders. Finally, it is important to highlight that existing studies were not powered to assess differences in pregnancy rates, rendering their use in counselling patients regarding treatment success rates uninformative. None of the differences observed in the included trials reached statistical significance with regard to clinical or ongoing pregnancy rates. This methodological coming is most probably due to the expected low pregnancy rates in poor responders (Surrey and Schoolcraft, 2000), such that sample sizes with enough power to detect differences in pregnancy rate would be prohibitive for a single centre trial. Recently, a

8 Article - Pituitary suppression s in poor responders - SK Sunkara et al. Cochrane Review has examined the various interventions in poor responders, including some of the comparisons made in the present review (Shanbhag et al., 2007). There were important differences in the inclusion and exclusion criteria between this review and the Cochrane Review, resulting in a difference in the selection of studies. However, the inferences were broadly similar, thus confirming the present findings. Because of these inconsistent findings, a well designed, adequately powered, multicentre three arm trial comparing the long versus the versus the ant is needed. Akman MA, Erden HF, Tosun SB et al Comparison of ic flare-up-protocol and antic multiple dose protocol in ovarian stimulation of poor responders: results of a prospective randomized trial. Human Reproduction 16, Akman MA, Erden HF, Tosun SB et al Addition of ant in cycles of poor responders undergoing IVF. Human Reproduction 15, Cheung LP, Lam PM, Lok IH et al ant versus long protocol in poor responders undergoing IVF: a randomized controlled trial. Human Reproduction 20, Crammer DW, Powers DR, Oskowitz SP et al Gonadotrophinreleasing hormone use in assisted reproduction cycles: the influence of long and s on pregnancy rates. Fertility and Sterility 72, CRD 2001 CRD Report Number 4, 2nd Edition. Undertaking Systematic Reviews of Research on Effectiveness. CRD s Guidance For Carrying Out of Commissioning Reviews. NHS Center for Reviews and Dissemination, University of York. Daya S 1997 Optimal protocol for gonadotrophin-releasing hormone use in ovarian stimulation. In: Gomel V, Cheung PCK (eds) In Vitro Fertilization and Assisted Reproduction. Monduzzi Editore, Bologna, pp De Placido G, Mollo A, Clarizia R et al Gonadotropin-releasing hormone () ant plus recombinant luteinizing hormone vs. a standard protocol in patients at risk for poor ovarian response. Fertility and Sterility 85, Dirnfeld M, Fruchter O, Yshai D et al Cessation of gonadotrophin-releasing hormone analogue (-a) upon downregulation versus conventional long -a protocol in poor responders undergoing in vitro fertilization. Fertility and Sterility 72, Faber BM, Mayer J, Cox B et al Cessation of gonadotropinreleasing hormone therapy combined with high-dose gonadotropin stimulation yields favorable pregnancy results in low responders. Fertility and Sterility 69, Garcia J, Padilla S, Bayati J, Baramki T 1990 Follicular phase gonadotrophin releasing hormone and human gonadotrophins: a better alternative for in vitro fertilization. Fertility and Sterility 53, Garcia-Velasco JA, Isaza V, Requena A et al High doses of gonadotrophins combined with stop versus non-stop protocol of analogue administration in low responder IVF patients: a prospective, randomized controlled trial. Human Reproduction 15, Griesinger G, Felberbaum RE, Schultze-Mosgau A, Diedrich K 2004 Gonadotropin-releasing hormone ants for assisted reproductive techniques. Are there clinical differences between agents? Drugs 64, Hugues J, Cedrin-Durnerin I 1998 Revisiting gonadotrophin-releasing hormone protocols and management of poor ovarian responses to gonadotrophins. Human Reproduction Update 4, Keay SD, Liversedge NH, Mathur RS, Jenkins JM 1997 Assisted conception following poor ovarian response to gonadotrophin stimulation. British Journal of Obstetrics and Gynaecology 104, Kenigsberg D, Littman BA, Hodgen GD 1984 Medical hypophysectomy: 1. Dose-response using a gonadotrophinreleasing hormone ant ganirelix for in vitro fertilization cycles in which the lead follicle is >14 mm. Fertility and Sterility 42, Klingmuller D, Schepke M, Enzweiler C, Bidlingmaier F 1993 Hormonal responses to new potent ant cetrorelix. Acta Endocrinologica 128, Lashen H, Ledger W 1999 Management of poor responders. Human Reproduction 14, Mahutte NG, Arici A 2002 Poor responders: does the protocol make a difference? Current Opinion in Obstetrics and Gynaecology 14, Malmusi S, La Marca A, Giulini S et al Comparison of a gonadotropin-releasing hormone () ant and flare-up in poor responders undergoing ovarian stimulation. Fertility and Sterility 84, Marci R, Caserta D, Dolo V, Tatone C et al ant in IVF poor-responder patients: results of a randomized trial. Reproductive BioMedicine Online 11, Morgia F, Sbracia M, Schimberni M et al A controlled trial of natural cycle versus microdose gonadotrophin-releasing hormone analog flare cycles in poor responders undergoing in vitro fertilization. Fertility and Sterility 81, Schmidt DW, Bremner T, Orris JJ et al A randomized prospective study of microdose leuprolide versus ganirelix in in vitro fertilization cycles for poor responders. Fertility and Sterility 83, Shanbhag S, Aucott L, Bhattacharya S et al Interventions for poor responders to controlled ovarian hyperstimulation (COH) in in-vitro fertilisation (IVF). Cochrane Database of Systematic Reviews. Issue1: Art No.: CD DOI: / CD pub2. Surrey ES, Schoolcraft WB 2003 Evaluating strategies for improving ovarian response of the poor responder undergoing assisted reproductive techniques. Fertility and Sterility 73, Weissman A, Farhi J, Royburt M et al Prospective evaluation of two stimulation protocols for low responders who were undergoing in vitro fertilization-embryo transfer. Fertility and Sterility 79, Declaration: The authors report no financial or commercial conflicts of interest. Received 16 May 2007; refereed 27 June 2007; accepted 7 August 2007.