Reproductive outcome after transplantation of ovarian tissue: a systematic review

Transcription

1 Human Reproduction Vol.23, No.12 pp , 2008 Advance Access publication on August 9, 2008 doi: /humrep/den301 Reproductive outcome after transplantation of ovarian tissue: a systematic review Mohamed A. Bedaiwy 1,2,3,, Sherif A. El-Nashar 3,, Ali M. El Saman 3, Johannes L.H. Evers 4, Samith Sandadi 2, Nina Desai 1 and Tommaso Falcone 1,5 1 Department of Gynecology and Obstetrics and Gynecology, the Cleveland Clinic Foundation, 9500 Euclid Avenue, Cleveland, OH 44195, USA; 2 University Hospitals of Cleveland, Cleveland, OH, USA; 3 Department of Obstetrics and Gynecology, Assiut University, Egypt; 4 Research Institute GROW, Maastricht University and Academisch ziekenhuis, Maastricht, The Netherlands 5 Correspondence address. falcont@ccf.org BACKGROUND: Despite interest in ovarian tissue transplantation (OTT) as a promising procedure for fertility preservation, to date, no precise data are available about its effectiveness. We systematically reviewed reproductive function after OTT for fertility preservation in women at high risk of premature ovarian failure (POF). METHODS: We searched the MEDLINE, EMBASE, Cochrane Systematic Reviews, CENTRAL, Web of Science and Scopus databases for studies on the reproductive outcomes after OTT in humans up to June Women with follicle-stimulating hormone (FSH) >30 IU/l at the time of OTT were included in a meta-analysis of individual-patient data to evaluate the time to re-establishment of ovarian function (ROF). Secondary outcomes included short-term (<12 months) and long-term (>12 months) ovarian function (OVF) and pregnancy after OTT. RESULTS: We identified 25 reports including 46 unique cases. OTT was performed to treat POF in 27 women, to prevent POF in 15, to treat infertility in 2 and accidentally in 1. In 23 women with FSH >30 at the time of OTT, OVF was re-established with a median time to ROF of 120 days (range ). Within 6 months after ROF, four women had recurrent ovarian failure. There are insufficient data to evaluate the long-term OVF (>12 months). Fresh grafts had an increased likelihood of return of OVF and a decreased likelihood for recurrent ovarian failure compared with cryopreserved grafts [HR of 2.44 (95% CI 0.92, 6.49) and 0.47 (95% CI 0.18, 1.12), respectively]. In 25 women who sought pregnancy, eight women had nine pregnancies at 12 months, giving a cumulative pregnancy rate of 37% (95% CI 19, 60). CONCLUSIONS: Transplantation of ovarian tissue can re-establish OVF after POF; however, the efficacy of OTT using cryopreserved tissues is not yet equivalent to that of fresh grafts. A controlled multicenter trial with sufficient follow-up would provide valid evidence of the potential benefit of this procedure. Keywords: systematic review; ovarian transplantation; ovarian function; pregnancy; cancer Introduction Ovarian tissue transplantation (OTT) is becoming an increasingly popular strategy for fertility preservation. The original indication was to restore fertility in cancer patients (autologous transplantation). Nevertheless, the spectrum has now expanded to incorporate OTT in twins discordant for premature ovarian failure (POF) and to restore ovarian function (OVP) in women with ovarian dysgenesis using ovarian tissue from matched donors, i.e. heterologous transplantation (Mhatre et al., 2005; Silber et al., 2005; Silber and Gosden, 2007). Another proposed indication, yet not pursued, is to prolong the reproductive life in otherwise healthy women (Silber and Gosden, 2007). Despite the experimental nature of the procedure, there is an increasing public awareness of its availability The first two authors participated equally in this work. as a potential fertility preservation strategy for those at risk. The American Society of Reproductive Medicine (ASRM) practice committee categorized the use of ovarian tissue cryopreservation with subsequent transplantation as an experimental option for fertility preservation with the notion that it should be performed after approval of the local Institutional Review Boards (Lee et al., 2006). Despite sparse evidence of its clinical usefulness, ovarian tissue banking has been offered to patients as a clinical service for almost two decades by many clinical centers around the globe (Grischenko et al., 1987; Martin et al., 2007; Poirot et al., 2007; Weintraub et al., 2007). The three potential uses proposed for harvested ovarian tissue were in vitro maturation of primordial follicles, xenografting of ovarian tissue or subsequent ovarian transplantation. To date, only the latter has been successfully attempted in humans (Oktay and Karlikaya, 2000; Silber and Gosden, 2007). # The Author Published by Oxford University Press on behalf of the European Society of Human Reproduction and Embryology. All rights reserved. For Permissions, please journals.permissions@oxfordjournals.org 2709

2 Bedaiwy et al. Successful attempts of OTT have been reported using a variety of ovarian tissue sizes ranging from cortical strips to the transplantation of the whole ovary with or without its vascular pedicle (Oktay, 2001; Oktay et al., 2001a; Donnez et al., 2006). Although fresh transplantation was possible for all these variety of sizes, transplantation of previously frozen ovarian tissues was only reported using cortical strips in humans. Most of the reported ovarian tissue cryopreservation protocols were modifications of the original one reported by Gosden et al. (1994). In addition, wide varieties of alternative sites and surgical techniques have been reported. Orthotopic ovarian transplantation has been tried to the ovarian stump or to the periovarian region (Oktay and Karlikaya, 2000; Donnez et al., 2004; Meirow et al., 2005; Silber and Gosden, 2007). Alternatively, heterotopic ovarian transplantation has been performed to a variety of locations including the arm (Leporrier et al., 1987; Callejo et al., 2001), the forearm (Oktay et al., 2001b; Wolner-Hanssen et al., 2005), rectus abdominis muscle (Kim et al., 2004), the subcutaneous tissue of the abdominal wall (Oktay et al., 2004) and the suprapubic area (Oktay, 2006). A combined orthotopic and heterotopic approach has been utilized as well (Schmidt et al., 2005; Rosendahl et al., 2006). The orthotopic transplantation process has been accomplished via either laparoscopy (Donnez et al., 2004) or laparotomy (Meirow et al., 2005; Silber and Gosden, 2007). Despite the relatively large number of reports on OTT approaches and techniques, to date, there are no precise data on the reproductive outcomes after OTT. This is in part due to the small numbers of patients and the wide range of indications, surgical techniques and duration of follow-up. The experimental nature as well as ethical concerns prevented the execution of larger well-designed studies to answer these important questions. We therefore performed a systematic review to summarize the reports on the reproductive outcomes after OTT for fertility preservation purposes, including prevention and reversal of POF. Materials and Methods Search strategy and identification of studies A systematic search on the MEDLINE, EMBASE, Cochrane Systematic Reviews, CENTRAL, Web of Science and Scopus databases from inception until June 2007 was performed. In this search, we used terms referring to ovarian tissue transplantation and reproductive outcomes. No language restriction was applied. Two reviewers (M.A.B. and S.A.E.) independently reviewed the identified reports. Disagreements in study inclusion were resolved by consensus and agreement in the selection of the studies was evaluated using kappa statistics. In addition, both reviewers independently reviewed the bibliographies of the retrieved articles and the recent reviews for additional studies. The search was conducted to include published peer reviewed studies on the reproductive outcomes after OTT in humans including re-establishment of ovarian function (ROF) evidenced by ovarian follicular growth or menstruation, and pregnancy. Outcomes The primary outcome for this review was the ROF after OTT as defined by evidence of follicular growth or the return of menstruation with the 2710 time to ROF as the primary end-point for this outcome. Secondary outcomes included short-term (,12 months) and long-term (.12 months) maintenance of OVF and the occurrence of pregnancy after the OTT procedure. Subgroup analyses were planned a priori and performed based on study characteristics (study design and the number of participants in each study), patients characteristics [etiology of OTT, age at POF, age at OTT, baseline follicle-stimulating hormone (FSH)], procedure-specific variables (fresh versus frozen, autologous versus heterologous and orthotopic versus heterotopic) and the amount of the transplanted tissues (whole ovary, whole or most of the cortex and cortical strips). We also reported on the pregnancy outcome in the subgroup of women who sought pregnancy including those who underwent in vitro fertilization (IVF). Data extraction A standardized data extraction sheet was developed and two reviewers (M.A.B. and S.A.E.) independently extracted the individual data from each included report. Disagreements were resolved by consensus. The methodological characteristics of the included studies including study design, number of participants and procedure-specific characteristics of OTT were reported. The baseline demographic and clinical individual patient-level data in the sheet included age at the time of OTT, indication for OTT, wish for pregnancy, medical and surgical history, documented occurrence of POF, the baseline FSH level and the age of the patient at POF, and previous chemotherapy or radiotherapy. The procedural-specific data included the technique and the approach used for OTT along with the amount of transplanted tissues. The duration of the follow-up was recorded in all patients. Authors were contacted to verify unclear information in a few instances (Silber and Gosden, 2007). Statistical analysis Count and percentages along with their exact binomial 95% confidence intervals (95% CIs) were reported for categorical data (Newcombe, 1998). Mean and standard deviation (SD) were used for continuous variables with normal distribution whereas median and inter-quartile range (25th percentile and the 75th percentile) were used for skewed data. Kaplan Meier method was used to estimate the cumulative rate for the primary and secondary outcomes and their corresponding 95% CI (Kaplan and Meier, 1958). Cox proportional hazard models were used to identify predictors of various outcomes after OTT and to explore sources of heterogeneity. For each outcome, univariate analyses were presented for clinical and procedural variables with potential impact on the outcome. A global (multivariate) analysis was then performed including selected unrelated co-variants along with a variable representing the included studies. Potential predictors were considered for inclusion in the global analysis if they had a P-value of,0.20 in the univariate analysis (Robins and Greenland, 1986). If zero events occurred in a subgroup, a correction of one event was added to each subgroup to make the calculation of hazard ratios defined. Sensitivity analyses were performed to evaluate the effect of this correction on variables included in the global analyses (Higgins and Green, 2008). Statistical analyses in this study were performed using RevMan software version (The Cochrane Collaboration, Oxford, UK) and JMP version 6.0 (SAS Institute Inc., Cary, NC, USA). Results Study identifications Out of 744 reports identified in the search, 25 reports met the inclusion criteria of the review (Fig. 1). All identified studies

3 Reproductive function after ovarian transplant ovaries were transplanted to their upper limb to avoid the effect of pelvic radiation as a treatment of in one patient and uterine cervical cancer in the other (Leporrier et al., 1987; Hilders et al., 2004). In the remaining patient, the transplant was performed to re-establish OVF in a patient with ovarian dysgenesis (Turner s syndrome). In that patient, a complete ovary with its vascular pedicle was removed form a HLA-matched donor (her 26-year-old sister) and transplanted into the pelvis with the ovarian vascular pedicle attached to the inferior epigastric vessels (Mhatre et al., 2005). In all three women, no documentations of ovarian failure by FSH elevation before ovarian transplant were reported. Nevertheless, OVF was reported in all three patients after transplant (Leporrier et al., 1987; Hilders et al., 2004; Mhatre et al., 2005). However, long-term OVF beyond 1 year and up to 18 years was only reported by Leporrier and colleagues (Leporrier et al., 2002a). Figure 1: Flow chart for selection of included reports. were observational studies including 3 cohort studies, 1 case series and 21 case reports. The total agreement in the selection of studies between the two reviewers was 95% with kappa of 0.9 (95% CI 0.8, 1.0). Out of these 25 reports, four reports included further technical and follow-up data about the same cases included in other studies by the same research groups (Leporrier et al., 1987; Von Theobald et al., 1987; Oktay et al., 2001b, 2003; Tryde Schmidt et al., 2004; Schmidt et al., 2005; Rosendahl et al., 2006). Finally, the review included 21 unique reports including 46 women who underwent transplantation of ovarian tissues from 1985 through 2007 (Table I). Clinical characteristics of the included studies Individual patient data was available for all 46 patients included. The median age was 36 years (IQR years, and range from years). OTT was performed to treat already diagnosed POF in 27 women, to avoid the development of POF in 16 women, to treat infertility in 2 women and in 1 woman evidence of return of OVF was documented after accidental entrapment of some ovarian tissues in the abdominal wall after laparoscopic oophorectomy for a benign indication. To date, only 3 out of 46 women had one or more whole ovaries freshly transplanted (no cryopreservation was performed) with or without microvascular attachment of their pedicle (Leporrier et al., 1987; Hilders et al., 2004; Mhatre et al., 2005). In two of these three women, the patients own Re-establishment of OVF (primary outcome) For the evaluation of the primary outcome (ROF), 27 out of the total identified 46 cases underwent OTT to re-establish their OVF after diagnosis of POF. Four of these women did not have documented FSH levels at the time of OTT and were excluded (Oktay and Karlikaya, 2000; Hilders et al., 2004; Mhatre et al., 2005). Thus, only 23 patients who had pre-ott FSH.30 IU/l were included in this analysis. Data needed for evaluation of the primary outcome were complete in all 23 patients (Table II). The return of OVF was based on the return of follicular growth and menstruation in 19 cases (18 spontaneous and 1 withdrawal bleeding), and detection of follicular growth in the remaining four cases (all had hysterectomies). The time needed to re-establish OVF ranged from 60 to 244 days with a median time to return of OVF of 120 days with all patients [100% (95% CI 91, 100)] showing at least some evidence of OVF at 244 days using Kaplan Meier Plot (Fig. 2). To explore heterogeneity, we compared the risk ratios for the return of OVF between different subgroups depending on patient and procedure characteristics. For procedure-related variables, fresh and heterologous OTT increased the likelihood of return of OVF [HR of 2.42 (95% CI 0.97, 6.04, P ¼ 0.058) and 9.30 (95% CI 2.45, 45.10, P ¼ 0.001), respectively]. In addition, larger sized grafts were more likely to reestablish OVF [HR of 4.67 (1.40, 15.04, P ¼ 0.012)]. In the global (multivariate) analysis, fresh grafts had an increased likelihood for ROF after OTT compared with cryopreserved grafts with an adjusted HR of 2.44 (95% CI 0.92, 6.49, P ¼ 0.073). Recurrent ovarian failure Data needed for evaluation of the time to recurrent ovarian failure were complete for all 23 patients; however, the median follow-up time after ROF was only 210 days (range days). At 6 months after the procedure, 4 out of 23 women had recurrence of ovarian failure with a cumulative recurrent failure rate of 22% (95%, 4 44%) (Fig. 3). This early recurrent failure was evidenced by increased FSH (.30 IU/l), reduced estrogen production and the failure of detection of follicular growth, despite induction with high 2711

4 2712 Table I. Methodological characteristics of the identified studies. Study Design POF Age at OTT Indication & Co-morbidities Need for future Pregnancy OTT Yes No Auto versus heterologous Fresh versus frozen Ortho versus heterotopic What was transplanted? Follow-up (month) Bedaiwy et al. 1 Leporrier et al. (1987) and Von Theobald et al. (1987) Report of 1 case To avoid POF after radiation Rx for 2 Marconi et al. (1997) Report of 1 case Accidental after Surgery for Rx of an endometrioma 3 Oktay (2000) and Oktay Report of 1 case To Rx POF after BSO for benign ov. et al. (2001a) tumor 4 Radford et al. (2001) Report of 1 case To Rx POF after chemotherapy for 5 Oktay et al. (2001b, 2003) Report of 2 cases and 37 Case1 to avoid POF after radiation Rx for cervical cancer IIIb 6 Callejo et al. (2001) Cohort (4 cases and 10 controls) Case2 to avoid POF after excision of benign ovarian tumor (46 49) To avoid POF in 3 and to Rx POF in 1 after hysterectomy-bso for fibroid uteri 7 Oktay et al. (2004) Report of 1 case To Rx POF after chemotherapy for breast cancer 8 Donnez et al. (2004) Report of 1 case To Rx POF after chemotherapy for 9 Kim et al. (2004) Report of 1 case To Rx POF after hysterectomy-bso for Cx cancer Ib 10 Kiran et al. (2004) Report of 1 case To Rx POF after hysterectomy-bso for fibroid uterus 11 Hilders et al. (2004) Report of 1 case To avoid POF after hysterectomy-bso for Cx cancer I-IIa 12 Wolner-Hanssen et al. (2005) Report of 1 case To Rx POF after chemotherapy and bone marrow transplant for Rx of Sjogren s with red cell aplasia Yes Auto Fresh Hetero Whole ovary 24 NA Auto Fresh Hetero Accidental entrapment of cortical fragments No Auto Cryo Ortho 80 cortical strips ( mm) Yes Auto Cryo Ortho 2 cortical strips ( mm) Both desired Auto Fresh Hetero Case1:16 cortical strip pregnancy with ( mm) IVF in Case1 Case2: whole cortex No Auto 3 Fresh and 1 cryo Hetero Whole cortex in 2 and 40 cortical strips in 2 (3 mm?) Yes (IVF) Auto Cryo Hetero 15 cortical strips ( mm) 8 Yes Auto Cryo Hetero 5 cortical strips 20 ( mm) No Auto Cryo Hetero 40 cortical strips 7 (5 5 1 mm) No Auto Fresh Hetero 10 cortical strips 6 ( mm) No Auto Fresh Hetero Whole ovary 12 Yes Auto Cryo Hetero 10 cortical strips (221 mm) 13 Mharte et al. (2005) Report of 2 cases and 15 To Rx POF in ov. dysgenesis (Turner s) Yes? Hetero Fresh Ortho Whole ovary in 1 and whole cortex in 1 14 Meirow et al. (2005) Report of 1 case To Rx POF after chemotherapy for Yes (IVF) Auto Cryo Ortho 6 cortical strips ( mm) 15 Donnez et al. (2005) Report of 2 cases and 27 To Rx infertility after removal of one Yes Auto Fresh Hetero?? 4 strips in 1 and 2 in the ovary other (10 4? mm) 16 Tryde Schmidt et al. (2004), Cohort (3 cases and (27 35) To Rx POF after chemotherapy for Yes (IVF) Auto Cryo Combined in 12 strips in 2 and 6 strips Schmidt et al. (2005), control) lymphoma (2 Hodgkin and 1 2 & ortho in 1 in 1 (5 5 1mm) Rosendahl et al. (2006) non-hodgkin) 17 Donnez et al. (2006) Report of 1 case To Rx POF after chemotherapy and bone marrow transplant for sickle cell anemia 18 Oktay (2006) Report of 1 case To Rx POF after chemotherapy for 19 Demeestere et al. (2006) Report of 1 case To Rx POF after chemotherapy for 20 Silber and Gosden (2007) Report of 7 cases (24 40) To Rx POF in ov. dysgenesis (in a series of monozygotic twins discordant for ov. failure) 21 Sanchez et al Cohort (12 cases and 5 controls) (37 45) To avoid POF in women undergoing hysterectomy-bso for benign Gynecological conditions (10 fibroids and 2 menorrhagia) OTT, ovarian tissue transplant; POF, premature ovarian failure; Rx, treatment; BSO, bilateral saplingo oopherectomy; ov, ovarian, Cx, cervix. No FSH level reported in these patients at or before the procedure. Ovarian transplant from HLA-matched sister (27 years) in one and from HLA-matched mother (38 years) in the second case. Yes Auto Cryo Combined 39 strips , , , 3 8 (8 22) Yes Auto Cryo Hetero Strips (number not reported) (5 5 1 mm) 11 Yes Auto Cryo combined 18 cortical strips 9 (5 5 2 mm) Yes (1 out of 7 Hetero Fresh Ortho 2/3 of the whole cortex 13 (7 28) had IVF) No Auto Fresh Ortho Whole cortex 24

5 Reproductive function after ovarian transplant Table II. Clinical and procedural characteristics of the women who underwent ovarian tissue transplant with FSH.30 IU/l at the time of transplantation. n ¼ 23 95% CI Etiology of POF Monozygotic twins discordant for POF 7 (30%) [17, 51%] Chemotherapy-induced POF 11 (48%) [29, 67%] Oopherectomy 5 (22%) [10, 42%] Age at POF, years [25.5, 32.9] Age at OTT, years [30.9, 36.1] Time from POF to OTT, years [2.0, 7.4] Age at OTT.35 years 12 (52%) [33, 71%] Pre OTT FSH Level, IU/l [69.7, 91.1] Fresh versus cryopreserved Fresh 10 (44%) [26, 63%] Cryopreserved 13 (56%) [37, 74%] Autologous versus Heterologous Autologous 16 (70%) [49, 84%] Heterologous 7 (30%) [17, 51%] Orthotopic versus Heterotopic Orthotopic 12 (52%) [33,71%] Heterotopic 7 (30%) [17,51%] Combined 4 (18%) [7, 37%] Ovarian tissue transplant size Whole Cortex 2 (9%) [2, 27%] 2/3 of the cortex 7 (30%) [17, 51%] Cortical strips 14 (61%) [41, 78%] Follow-up, days [311.0, 478.1] Return of follicular growth 23 (100%) [86, 100%] Return of menstruation 19 (83%) [63, 93%] Recurrent ovarian failure during the 4 (18%) [7, 37%] follow-up period Pregnancy 8 (40%) [22, 61%] 95% CI, 95% confidence interval; POF, premature ovarian failure; OTT, ovarian tissue transplantation; FSH, follicle-stimulating hormone. doses of FSH in all four patients. Long-term evaluation of OVF was not possible due to the insufficient length of follow-up in the published reports. Follow-up of OVF beyond 12 months was only available for eight cases (38%) (Oktay et al., 2001b; Schmidt et al., 2005; Rosendahl et al., 2006; Silber and Gosden, 2007). In these eight patients, the re-established OVF was documented till the end of the follow-up. Women who had recurrent ovarian failure were significantly older at OTT with a median age of 37.5 years (range 35 47; Figure 2: Time to the return of ovarian function and the corresponding 95% confidence interval after OTT (n ¼ 23). Figure 3: Cumulative rate for early recurrent ovarian failure in women who reestablished ovarian function after OTT (n ¼ 23). The dashed lines are the 95% CI. IQR 35.5, 44.8) compared with 33 years (range 24, 44; IQR 27, 36) in those who did not have recurrent ovarian failure [HR of 1.18 (95% CI 1.01, 1.14, P ¼ 0.035)]. The median age at which the patient was diagnosed with POF prior to transplant was 36.5 years (range 30, 47) compared with 28 years (range 13, 44) in those who did not have recurrent ovarian failure after [HR of 1.15 (95% CI 1.02, 1.30), P ¼ 0.022]. In addition, all four women who had recurrent failure had cryopreserved autologous, grafts which consisted of smaller cortical strips. The four women did re-establish their OVF after 90 days. In the global (multivariate) analysis, fresh grafts had a decreased likelihood for recurrent ovarian failure after OTT compared with cryopreserved grafts with adjusted HR of 0.47 (95% CI 0.18, 1.12, P ¼ 0.090) (Global chi-square estimate for this multivariate model was with three degrees of freedom and P ¼ 0.018). Pregnancy rate after ovarian transplant For the evaluation of pregnancy after ovarian transplant, out of the identified 46 cases included in the review, 21 women were excluded from this part of the analysis as they had hysterectomy or were reported by the authors that they did not seek pregnancy (Marconi et al., 1997; Callejo et al., 2001; Hilders et al., 2004; Kim et al., 2004; Kiran et al., 2004; Sanchez et al., 2007). Thus, 25 women with intact uteri were included in the final analysis for pregnancy after OTT. These women included 20 women with evidence of POF before OTT (FSH.30 IU/l) (Radford et al., 2001; Donnez et al., 2004; Oktay et al., 2004; Meirow et al., 2005; Schmidt et al., 2005; Wolner- Hanssen et al., 2005; Demeestere et al., 2006; Donnez et al., 2006; Oktay, 2006; Rosendahl et al., 2006; Silber and Gosden, 2007), two women with endometriosis who underwent OTT for preservation of ovarian tissue from an ovary that was not possible to conserve (Donnez et al., 2005), two young women with Turner s syndrome who underwent heterologous fresh orthotopic OTT using partial cortex in one and the whole ovary in the other (Mhatre et al., 2005), and the last patient who underwent chemotherapy for Hodgkin s but did not have evidence of ovarian failure at the time of OTT 2713

6 Bedaiwy et al. (Leporrier et al., 1987). In these 25 women, a total of 11 pregnancies in nine women were reported. However, one of these women became pregnant spontaneously twice after heterotopic OTT in the suprapubic area of the abdominal wall without performing IVF and was censored from the analysis at the time of her first pregnancy (Oktay, 2006). Thus, the final analysis included a total of nine pregnancies in eight women in five reports (Donnez et al., 2004; Meirow et al., 2005; Demeestere et al., 2006; Rosendahl et al., 2006; Silber and Gosden, 2007). The 12 month cumulative pregnancy rate was 37% (95% CI 19 60%) (Fig. 4). In those eight women; the first pregnancy was achieved after a median duration of 9 months after OTT (range months). In one woman, a second pregnancy was achieved 25.5 months after OTT and in the eighth menstrual cycle after delivery (Silber and Gosden, 2007). In four out of those eight women, autologous OTT using cryopreserved ovarian tissues was performed (Donnez et al., 2004; Meirow et al., 2005; Demeestere et al., 2006; Rosendahl et al., 2006), whereas in the remaining four women, heterologous OTT using fresh ovarian tissue from their monozygotic twins discordant for ovarian failure was performed (Silber and Gosden, 2007). To date, no pregnancies have been reported after the three published cases of whole intact fresh human ovary transplantation. Although all three had intact uteri, none has yet pursued pregnancy (Leporrier et al., 1987; Hilders et al., 2004; Mhatre et al., 2005). To explore potential heterogeneity in the pregnancy outcome among the included studies, we performed a stratified analysis based on study and patient characteristics. In the global analysis, only lower OTT FSH level was a significant predictor of pregnancy after OTT with adjusted HR of 1.06 (95% CI 1.01, 1.13, P ¼ 0.011) (The global chi-square for this multivariate model was 9.37 with four degrees of freedom and P ¼ ) (Table III). Pregnancies were achieved spontaneously in five women (six pregnancies), whereas in the remaining three women, pregnancies were achieved through IVF (natural cycle IVF in two and antagonist IVF protocol in one). The diagnosis of pregnancy was biochemical in one patient (Rosendahl et al., Figure 4: Cumulative pregnancy rate in women who sought pregnancy after OTT (n ¼ 25). The dashed lines are the 95% CI. 2006), whereas in the other seven women, clinical pregnancy was diagnosed based on the detection of gestational sac and embryonic pole by ultrasound assessment (Donnez et al., 2004; Meirow et al., 2005; Demeestere et al., 2006; Silber and Gosden, 2007) (Table IV). Discussion An important finding of this study is that regardless of the technique used or the approach of OTT, some degree of OVF was restored in all patients. This supports the short-term efficacy of the procedure in humans. Another significant finding is that the efficacy of OTT using cryopreserved tissues is not yet equivalent to that of fresh grafts. This warrants for further improvements in cryopreservation protocols and surgical techniques to optimize the outcome. OVF after cryopreserved compared with fresh ovarian transplant The increased likelihood of return of OVF in fresh grafts compared with cryopreserved grafts can be explained by the relatively larger cortical tissue size in fresh grafts resulting in transplantation of a larger cohort of viable follicles compared with the cryopreserved grafts (Silber and Gosden, 2007). Given the fact that previous animal studies suggested that cryopreservation of ovarian tissues does not substantially change the reproductive outcome after OTT, this observation could not be explained solely by the possible negative effects of the cryopreservation protocol (Aubard et al., 1999; Baird et al., 1999; Candy et al., 2000). However, further improvement in the cryopreservation techniques is still required to optimize the outcome after OTT of cryopreserved ovarian tissues in humans (Torrents et al., 2003; Bedaiwy et al., 2006; Bedaiwy and Falcone, 2007). Delayed ROF.90 days was significantly associated with a higher likelihood for early recurrent ovarian failure. This is most likely the result of the warm ischemic insult after OTT. Consequently, ovarian transplants with a vascular pedicle are currently being explored to ensure immediate graft revascularization, possible reduction in ovarian failure recurrence and prolongation of graft survival (Martinez-Madrid et al., 2004; Bedaiwy et al., 2006; Jadoul et al., 2007). The site of OTT whether orthotopic or heterotopic did not seem to affect the return of OVF or recurrent ovarian failure. Additionally, the age at OTT of the patient also did not affect the likelihood of return of OVF; however, older women at OTT were more likely to have recurrent ovarian failure. To date, there is insufficient evidence to support the longterm efficacy of OTT after fresh and cryopreserved OTT. For whole ovarian transplant, the first case, reported by Leporrier in 1987, was reported to still have evidence of OVF in 2002 (Leporrier et al., 1987, 2002b). However, this patient has her other ovary in place and was reported to have follicular growth in her pelvic ovary 6 months after OTT. In patients with evidence of POF at the time of OTT, only eight had follow-up.12 months and for studies reporting on cryopreserved OTT, only four women had.12 months follow-up after the ROF (Oktay et al., 2001b; Tryde Schmidt et al., 2714

7 Reproductive function after ovarian transplant Table III. Factors affecting the likelihood of pregnancy after OTT*. Univariate Multivariate (global) Hazard Ratio (95% CI) P-value Adjusted hazard ratio (95% CI) P-value Age at OTT, per year 0.98 (0.89, 1.08) Age at POF, per year 0.94 (0.85, 1.03) (0.74, 1.06) Pre OTT FSH Level, per IU/l 1.02 (0.99, 1.05) (1.01, 1.13) Fresh versus cryopreserved 1.09 (0.26, 4.64) (0.41, 52.64) Heterologous versus autologous 3.53 (0.77, 18.06) Orthotopic versus heterotopic 1.96 (0.32, 37.60) Size of tissue transplanted (whole or 2/3 of the cortex) versus cortical strips E 1.66 (0.19, 22.92) Study g (1.56, ) (0.89, 1.20) % CI, 95% confidence interval; POF, premature ovarian failure; OTT, ovarian tissue transplantation; FSH, follicle-stimulating hormone. *The multivariate global analysis included a variable representing the included studies. Because all fresh grafts were of larger size and most of them were heterologous grafts, we selected to include the variable Fresh versus Cryopreserved in global analysis. The global chi-square for the model was 9.37 with four degrees of freedom and P ¼ E In a sensitivity analysis for the size of the transplanted ovarian tissues, when considering the two cases (Donnez et al., 2004, 2006) as larger grafts, the univariate HR was 2.57 (95% CI, 1.04, 6.50); however, no effect on the final global analyses was noticed. g A variable related to the included studies. Table IV. Outcome of 20 IVF cycles in eight patients after OTT. n ¼ 20 Type of stimulation IVF cycles Natural cycle protocol 6/20 (30%) Antagonist protocol 9/20 (45%) Agonist protocol 5/20 (25%) Duration of COS, days 10.5 (range 2 17) Dose of FSH, Units (range ) Number of follicles at the day of hcg 1.00 (range 0 4) Number retrieved oocyctes 1.00 (range 0 4) Canceled cycle (no embryo transfer) 14/20 [70% (95% CI 46, 88%)] Failed retrieval 5/14 (36%) Failed or abnormal Fertilization 8/14 (57%) Failed embryonic development 1/14 (7%) Completed cycles (þembryo transfer) 6/20 [30% (95% CI 12, 54%)] Total pregnancies 3 pregnancies and only 1 delivery Chemical pregnancy 1/3 (33.3%) Ongoing pregnancy 1/3 (33.3%) Full-term delivery 1/3 (33.3%) Pregnancy rate per patient 3/8 [37.5% (95% CI 9, 76%)] Pregnancy per cycle started 3/20 [15% (95% CI 3, 38%)] Pregnancy per cycle completed 3/6 [50% (95% CI 12, 88%)] CI 95%, 95% confidence interval; IVF, in vitro fertilization; COS, controlled ovarian stimulation, FSH, follicle-stimulating hormone. 95% CI s were calculated using exact methods of a binomial parameter. 2004; Rosendahl et al., 2006). Despite the fact that at the end of the follow-up period, each of these four cases who underwent cryopreserved OTT had evidence of OVF, the overall recurrent ovarian failure was documented in.20% of the cases. Pregnancy after cryopreserved compared with fresh ovarian transplant Pregnancy is an important reproductive outcome after OTT, and for some groups of women, it represents one of the main goals of the technique. To date, 25 women who had OTT sought pregnancy and a total of nine pregnancies from the transplanted ovarian tissues in eight women were documented. To date, three women gave birth to three full-term babies and four women still have ongoing pregnancies. Two deliveries were after autotransplantation of cryopreserved-thawed ovarian tissue (Donnez et al., 2004; Meirow et al., 2005) and the third was after heterologous transplantation of fresh cortical tissue between twins discordant for POF (Silber et al., 2005). An interesting finding is the comparable likelihood of pregnancy after cryopreserved and fresh OTT. This finding supports previous results from animal studies suggesting similar reproductive outcomes using cryopreserved and fresh grafts (Aubard et al., 1999; Baird et al., 1999; Candy et al., 2000). One important concern is the source of oocytes in pregnancies after OTT. The source of oocytes was verified in the case of heterotopic OTT when fertilization and chemical pregnanacy resulted from an oocyte aspirated from the heterotopic transplanted ovarian tissue (Rosendahl et al., 2006). Despite the high likelihood of ovarian failure in young women undergoing chemotherapy and radiotherapy for treatment of cancer, spontaneous restoration of OVF after documented POF was documented in two published reports. In the first, spontaneous pregnancy was reported in a young woman 6 years after receiving combined chemotherapy and radiation for treatment for Ewing s sarcoma without receiving any ovarian transplants (Bath et al., 2004) and in the second report, pregnancy occurred in a young woman 2.5 years after chemotherapy for Hodgkin s lymphoma (Oktay, 2006). Nevertheless, the possibility of pregnancy after treatment of cancer is by itself an important step that provides hope for patients with cancer. Limitations This systematic review was limited by the constraints of the published data and small number of patients and limited follow-up, as well as by the fact that data are either case reports or small case series. Publication bias is an important threat to the validity of the review, especially with the inclusion of case reports. In addition, almost all the available reports are from experts, and wider scale use of this technique is expected to result in less favorable outcomes. Thus, caution is recommended in the interpretation of the success rates of the procedure in wider scale. Another important limitation is the heterogeneity in the 2715

8 Bedaiwy et al. methods and surgical approach of the operation as well as the various indications. To limit the effect of heterogeneity on the primary outcome, we only included patients with robust evidence of ovarian failure before the procedure to decrease the heterogeneity at baseline. The absence of a control group in these studies is another important limitation. In conclusion, transplantation of ovarian tissue can re-establish OVF after POF; however, evidence is not sufficient to evaluate the long-term efficacy and longevity of ovarian grafts. Future research should be directed toward optimization of the technology utilized in cryopreservation of ovarian tissue and in the standardization of the laboratory and surgical technology used (Torrents et al., 2003) as well as the use of vascularized grafts to minimize ischemic damage (Bedaiwy et al., 2006). After optimization of the technique, the next step would be to conduct a multicenter controlled clinical trial with sufficient follow-up (5 years) to systematically evaluate the efficacy of this procedure before wide clinical utilization. The sample needed for such a study is expected to be small, given the previously reported high efficacy of ovarian transplantation. The suggested primary end-point would be the rate of recurrent ovarian failure and pregnancy after OTT. Stratification based on the type of cancer, the chemotherapy regimen used and other relevant patient and procedural characteristics is recommended. Acknowledgement The Authors would like to thank Dr David Starks from the Department of Gynecology and Obstetrics and Gynecology, University Hospitals of Cleveland, Cleveland, OH, USA, for his assistance. References Aubard Y, Piver P, Cogni Y, Fermeaux V, Poulin N, Driancourt MA. Orthotopic and heterotopic autografts of frozen-thawed ovarian cortex in sheep. Hum Reprod 1999;14: Baird DT, Webb R, Campbell BK, Harkness LM, Gosden RG. Long-term ovarian function in sheep after ovariectomy and transplantation of autografts stored at 2196ºC. Endocrinology 1999;140: Bath LE, Tydeman G, Critchley HO, Anderson RA, Baird DT, Wallace WH. Spontaneous conception in a young woman who had ovarian cortical tissue cryopreserved before chemotherapy and radiotherapy for a ewing s sarcoma of the pelvis: case report. Hum Reprod 2004;19: Bedaiwy MA, Falcone T. Harvesting and autotransplantation of vascularized ovarian grafts: approaches and techniques. Reprod Biomed Online 2007;14: Bedaiwy MA, Hussein MR, Biscotti C, Falcone T. Cryopreservation of intact human ovary with its vascular pedicle. Hum Reprod 2006;21: Callejo J, Salvador C, Miralles A, Vilaseca S, Lailla JM, Balasch J. Long-term ovarian function evaluation after autografting by implantation with fresh and frozen-thawed human ovarian tissue. J Clin Endocrinol Metab 2001;86: Candy CJ, Wood MJ, Whittingham DG. Restoration of a normal reproductive lifespan after grafting of cryopreserved mouse ovaries. Hum Reprod 2000;15: Demeestere I, Simon P, Buxant F, Robin V, Fernandez SA, Centner J, Delbaere A, Englert Y. Ovarian function and spontaneous pregnancy after combined heterotopic and orthotopic cryopreserved ovarian tissue transplantation in a patient previously treated with bone marrow transplantation: case report. Hum Reprod 2006;21: Donnez J, Dolmans MM, Demylle D, Jadoul P, Pirard C, Squifflet J, Martinez-Madrid B, van Langendonckt A. Livebirth after orthotopic transplantation of cryopreserved ovarian tissue. Lancet 2004;364: Donnez J, Squifflet J, Dolmans MM, Martinez-Madrid B, Jadoul P, Van Langendonckt A. Orthotopic transplantation of fresh ovarian cortex: a report of two cases. Fertil Steril 2005;84:1018. Donnez J, Dolmans MM, Demylle D, Jadoul P, Pirard C, Squifflet J, Martinez-Madrid B, Van Langendonckt A. Restoration of ovarian function after orthotopic (intraovarian and periovarian) transplantation of cryopreserved ovarian tissue in a woman treated by bone marrow transplantation for sickle cell anaemia: case report. Hum Reprod 2006;21: Gosden RG, Baird DT, Wade JC, Webb R. Restoration of fertility to oophorectomized sheep by ovarian autografts stored at 2196 degrees C. Hum Reprod 1994;9: Grischenko V, Chub N, Lobyntseva G. Creation of bank of cryopreserved human ovarian for allotransplantation in gynecology. Kriobiologia 1987;3: Higgins J, Green S (eds). Cochrane Handbook for Systematic Reviews of Interventions (updated February 2008). The Cochrane Collaboration, Hilders CG, Baranski AG, Peters L, Ramkhelawan A, Trimbos JB. Successful human ovarian autotransplantation to the upper arm. Cancer 2004;101: Jadoul P, Donnez J, Dolmans MM, Squifflet J, Lengele B, Martinez-Madrid B. Laparoscopic ovariectomy for whole human ovary cryopreservation: technical aspects. Fertil Steril 2007;87: Kaplan E, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc 1958;153: Kim SS, Hwang IT, Lee HC. Heterotopic autotransplantation of cryobanked human ovarian tissue as a strategy to restore ovarian function. Fertil Steril 2004;82: Kiran G, Kiran H, Coban YK, Guven AM, Yuksel M. Fresh autologous transplantation of ovarian cortical strips to the anterior abdominal wall at the pfannenstiel incision site. Fertil Steril 2004;82: Lee SJ, Schover LR, Partridge AH, Patrizio P, Wallace WH, Hagerty K, Beck LN, Brennan LV, Oktay K. American society of clinical oncology recommendations on fertility preservation in cancer patients. J Clin Oncol 2006;24: Leporrier M, von Theobald P, Roffe JL, Muller G. A new technique to protect ovarian function before pelvic irradiation. Heterotopic ovarian autotransplantation. Cancer 1987;60: Leporrier M, Roffe JL, Von Theobald P, Muller G. Autologous transplantation of whole ovaries vs ovarian cortical strips (comment). JAMA 2002a;287: Leporrier M, Roffe JL, Von Theobald P, Muller G. Autologous transplantation of whole ovaries vs ovarian cortical strips. JAMA 2002b;287: Marconi G, Quintana R, Rueda-Leverone NG, Vighi S. Accidental ovarian autograft after a laparoscopic surgery: case report. Fertil Steril 1997;68: Martin JR, Kodaman P, Oktay K, Taylor HS. Ovarian cryopreservation with transposition of a contralateral ovary: a combined approach for fertility preservation in women receiving pelvic radiation. Fertil Steril 2007;87: 189 e5 189 e7. Martinez-Madrid B, Dolmans MM, Van Langendonckt A, Defrere S, Donnez J. Freeze-thawing intact human ovary with its vascular pedicle with a passive cooling device. Fertil Steril 2004;82: Meirow D, Levron J, Eldar-Geva T, Hardan I, Fridman E, Zalel Y, Schiff E, Dor J. Pregnancy after transplantation of cryopreserved ovarian tissue in a patient with ovarian failure after chemotherapy. N Engl J Med 2005;353: Mhatre P, Mhatre J, Magotra R. Ovarian transplant: a new frontier. Transplant Proc 2005;37: Newcombe RG. Two-sided confidence intervals for the single proportion: comparison of seven methods. Stat Med 1998;17: Oktay K. Ovarian tissue cryopreservation and transplantation: preliminary findings and implications for cancer patients. Hum Reprod Update 2001;7: Oktay K. Spontaneous conceptions and live birth after heterotopic ovarian transplantation: is there a germline stem cell connection? Hum Reprod 2006;21: Oktay K, Karlikaya G. Ovarian function after transplantation of frozen, banked autologous ovarian tissue. N Engl J Med 2000;342:1919. Oktay K, Aydin BA, Karlikaya G. A technique for laparoscopic transplantation of frozen-banked ovarian tissue. Fertil Steril 2001a;75: Oktay K, Economos K, Kan M, Rucinski J, Veeck L, Rosenwaks Z. Endocrine function and oocyte retrieval after autologous transplantation of ovarian cortical strips to the forearm. JAMA 2001b;286:

9 Reproductive function after ovarian transplant Oktay K, Buyuk E, Rosenwaks Z, Rucinski J. A technique for transplantation of ovarian cortical strips to the forearm. Fertil Steril 2003;80: Oktay K, Buyuk E, Veeck L, Zaninovic N, Xu K, Takeuchi T, Opsahl M, Rosenwaks Z. Embryo development after heterotopic transplantation of cryopreserved ovarian tissue. Lancet 2004;363: Poirot CJ, Martelli H, Genestie C, Golmard JL, Valteau-Couanet D, Helardot P, Pacquement H, Sauvat F, Tabone MD, Philippe-Chomette P et al. Feasibility of ovarian tissue cryopreservation for prepubertal females with cancer. Pediatr Blood Cancer 2007;49: Radford JA, Lieberman BA, Brison DR, Smith AR, Critchlow JD, Russell SA, Watson AJ, Clayton JA, Harris M, Gosden RG et al. Orthotopic reimplantation of cryopreserved ovarian cortical strips after high-dose chemotherapy for hodgkin s lymphoma. Lancet 2001;357: Robins JM, Greenland S. The role of model selection in causal inference from nonexperimental data. Am J Epidemiol 1986;123: Rosendahl M, Loft A, Byskov AG, Ziebe S, Schmidt KT, Andersen AN, Ottosen C, Andersen CY. Biochemical pregnancy after fertilization of an oocyte aspirated from a heterotopic autotransplant of cryopreserved ovarian tissue: case report. Hum Reprod 2006;21: Sanchez M, Alama P, Gadea B, Soares SR, Simon C, Pellicer A. Fresh human orthotopic ovarian cortex transplantation: long-term results. Hum Reprod 2007;22: Schmidt KL, Andersen CY, Loft A, Byskov AG, Ernst E, Andersen AN. Follow-up of ovarian function post-chemotherapy following ovarian cryopreservation and transplantation. Hum Reprod 2005;20: Silber SJ, Gosden RG. Ovarian transplantation in a series of monozygotic twins discordant for ovarian failure. N Engl J Med 2007;356: Silber SJ, Lenahan KM, Levine DJ, Pineda JA, Gorman KS, Friez MJ, Crawford EC, Gosden RG. Ovarian transplantation between monozygotic twins discordant for premature ovarian failure. N Engl J Med 2005;353: Torrents E, Boiso I, Barri PN, Veiga A. Applications of ovarian tissue transplantation in experimental biology and medicine. Hum Reprod Update 2003;9: Tryde Schmidt KL, Yding Andersen C, Starup J, Loft A, Byskov AG, Nyboe Andersen A. Orthotopic autotransplantation of cryopreserved ovarian tissue to a woman cured of cancer follicular growth, steroid production and oocyte retrieval. Reprod Biomed Online 2004;8: Von Theobald P, Roffe JL, Berrocal J, Le Porrier M, Levy G, Muller G. Heterotopic autotransplantation of the ovary in women. Presse Med 1987;16: Weintraub M, Gross E, Kadari A, Ravitsky V, Safran A, Laufer N, Revel A. Should ovarian cryopreservation be offered to girls with cancer. Pediatr Blood Cancer 2007;48:4 9. Wolner-Hanssen P, Hagglund L, Ploman F, Ramirez A, Manthorpe R, Thuring A. Autotransplantation of cryopreserved ovarian tissue to the right forearm 4(1/2) years after autologous stem cell transplantation. Acta Obstet Gynecol Scand 2005;84: Submitted on February 3, 2008; resubmitted on May 31, 2008; accepted on June 6,