The German Middleway as Precursor for Single Embryo Transfer. A Retrospective Data-analysis of the Düsseldorf University Hospital's Interdisciplinary Fertility Centre - UniKiD.

!

Patients receiving fertility treat- ment in Germany appear to be disadvantaged in comparison to those in other countries due to the restrictive Embryo Protection Act ( “Embry- onenschutzgesetz, ESchG ”), which prohibits the selection of a “top” embryo. The so-called Ger- man Middleway ( “Deutscher Mittelweg, DMW ”) now provides for a liberal interpretation of the ESchG by allowing the culture of numerous pro- nuclear stages (2PN stage).

Retrospective cohort study of 2 assisted reproduction treatment cycles in n = 400 patients between the ages of 21 and 45 years, either treated 2× conservatively or 1× con- servatively and 1× liberally according to DMW.

Pregnancy was achieved in 35 % of pa- tients in the DMW group and 31 % of controls. The birth rate among controls was 28.5 % and 30.5 % in the DMW group. Most pregnancies re- sulted from the culture of 4 × 2PN stages.

Patients in the DMW group had sig- nificantly higher pregnancy and birth rates com- pared to their previous cycles despite signifi- cantly increased age and significantly fewer transferred embryos. Key factors were the num- ber of 2PNs generated and the quality of embryos transferred. Thus it can be assumed that particu- larly older patients with adequate ovarian re- serves will benefit from DMW, i.e. the transfer of fewer embryos of the best possible quality. Zusammenfassung ! Einleitung: Deutsche Kinderwunschpatienten er- scheinen im internationalen Vergleich aufgrund der restriktiven Vorgaben des Embryonenschutz- gesetzes (ESchG) und des Verbots der Auswahl eines „Top-Embryos“ benachteiligt. Die Durch- führung des Deutschen Mittelwegs (DMW) er- laubt nun eine liberale Interpretation des ESchG im Sinne der Kultur mehrerer Vorkernstadien (2PN-Stadien).

und Methoden: Retrospektive Kohor- tenstudie mit 2 Behandlungszyklen im Rahmen einer assistierten Reproduktion mit n = 400 Pa- tientinnen im Alter zwischen 21 und 45 Jahren, die entweder 2 × konservativ oder 1 × konservativ und 1 × liberal nach DMW behandelt wurden. Ergebnisse: Eine Schwangerschaft konnte in der DMW-Gruppe in 35 % der Fälle, in der Kontroll- gruppe in 31 % erzielt werden. Die Geburtenrate lag in der Kontrollgruppe bei 28,5 % und in der DMW-Gruppe bei 30,5 %. Die meisten Schwanger- schaften resultierten aus der Kultur von 4 × 2PN- Stadien. Schlussfolgerung: Im Vergleich zum Vorzyklus einer Patientin der DMW-Gruppe kam es zu sig- nifikant höheren Schwangerschafts- und Gebur- tenraten bei gleichzeitig signifikant gestiegenem Alter und signifikant weniger transferierten Em- bryonen. Entscheidende Einflussfaktoren waren die Anzahl der generierten 2PN-Stadien und die Qualität der transferierten Embryonen. Somit ist davon auszugehen, dass gerade ältere Patientin- nen mit einer ausreichenden ovariellen Reserve vom DMW bzw. vom Transfer von weniger Em- bryonen mit möglichst hoher Qualität profitieren. The German Middleway as Precursor for Single Embryo Transfer. A Retrospective Data-analysis of the Düsseldorf University Hospital ʼs Interdisciplinary Fertility Centre – UniKiD Ebnet der DMW den Weg zum Single Embryo Transfer? Eine retrospektive Datenerhebung des universitären interdisziplinären Kinderwunschzentrums Düsseldorf – UniKiD Authors T. K. Kliebisch 1, A. P. Bielfeld 2, J. S. Krüssel 2, D. M. Baston-Büst 2 Affiliations 1 Medical Research School der Medizinischen Fakultät der Heinrich-Heine-Universität, Düsseldorf 2 Medical Center University of Düsseldorf, Dept. of OB/GYN/REI, Düsseldorf Key words l" ART l" German Middleway l" embryo protection act l" blastocyst Schlüsselwörter l" ART l" Deutscher Mittelweg l" ESchG l" Blastozyste received 4. 1. 2016 revised 9. 3. 2016 accepted 29. 3. 2016

DOI http://dx.doi.org/ 10.1055/s-0042-105747 Geburtsh Frauenheilk 2016; 76: 690–698 © Georg Thieme Verlag KG Stuttgart · New York · ISSN 0016‑5751 Correspondence Dunja Maria Baston-Büst Universitätsklinikum der Heinrich-Heine-Universität Frauenklinik Universitäres interdisziplinäres Kinderwunschzentrum Düsseldorf (UniKiD) Moorenstraße 5 40225 Düsseldorf [email protected] 690 Kliebisch TK et al. The German Middleway … Geburtsh Frauenheilk 2016; 76: 690 –698 Deutsche Version unter: http://dx.doi.org/ 10.1055/s-0042-105747 GebFra Science

! Approximately 10 % of couples of reproductive age are affected by involuntary childlessness [1]. Reproductive medicine in Germany is more strictly regulated than other countries in Europe and worldwide in terms of exploring new fertility treatment methods through national statutory framework conditions such as the Embryo Protection Act (ESchG) and some regional medical coun- cil professional codes of conduct. In other countries such as Bel- gium and Sweden elective single embryo transfer (eSET), where a single embryo is selected from those generated in a particular stimulation cycle and reimplanted, is legally sanctioned and in fact regarded as desirable by the health sector [2]. The 1990 German ESchG does not allow stock fertilisation, avoid- ing the regular production of surplus embryos [3]. Thus in Ger- many, during a single treatment cycle not all harvested ova are cultured, but rather, in the southern provinces since approx. 2008/2009 the so-called German Middleway (DMW) has been performed whereby, according to the couples wishes, as many ova as necessary are cultivated beyond the 2 pronuclear stage (2PN stage) to enable identification of 2 viable embryos for selec- tion [4]. Many reproductive medicine centres still cultivate ac- cording to the classical so-called “rule of three ”, which allows a maximum of 3 × 2PN stages, since no more than 3 embryos are allowed to be transferred per cycle [3]. From the legal perspec- tive, despite this liberal interpretation of the ESchG, conditions are not the same in Germany as in many neighbouring European countries where doctors often cultivate all fertilised ova for sub- sequent targeted selection of a single, optimally developed em- bryo for transfer. Through the selection of a best or “top” embryo in most cases surplus embryos are produced, meaning that this so-called elective Single Embryo Transfer (eSET) is not permitted in Germany [5]. This situation is the result of the 1991 ESchG, which stipulates that in order to avoid surplus embryos no stock fertilisation is permissible [3]. Current data however clearly show that only 1 – 2 viable embryos are produced on culture of up to 6 × 2PNs in 85 % of observed stimulation cycles, indicating that the concern surrounding embryo stockpiling may be exaggerated [6]. When one considers the three most important criteria for preg- nancy occurrence, i.e. patient age, number of transferred em- bryos and the embryo “quality”/viability, it is apparent that Ger- man fertility practitioners only achieve good pregnancy rates through transferring multiple embryos, which carries the risk of multiple pregnancy. The annual report of the German IVF Register (DIR) from 2013 showed that this risk is particularly high among 30 – 34 year olds following transfer of two embryos: between 2000 and 2012 20.1 % of births in this group were twins and 0.41 % triplets [7]. The DMW (so-called German Middleway) attempts to improve treatment results in Germany through a liberal interpretation of the ESchG which, after in vitro culture of numerous embryos and selection of those with the greatest potential viability, allows transfer at the latest at the blastocyst stage possibly even as a sin- gle embryo transfer depending on circumstances. This article outlines to what extent the DMV differs from stan- dard artificial insemination cycles, and whether its introduction at the UniKiD in 2011 has lead to significantly higher pregnancy rates, fewer multiple pregnancies and consequently significant reductions in health and financial burdens of treated couples. We pay particular attention to the question of which patient groups are able to benefit from DMW, in other words, which management algorithms predict success with this method. This is especially important to clarify whether DMW, as implemented in UniKiD and most other centres in Nordrhein as well as in many centres in other regional medical council areas, will pave the way towards Single Embryo Transfer.

! Procedure description The indication for reproductive medical treatment (in vitro fertil- isation [IVF] or intracytoplasmic sperm injection [ICSI]) was made after comprehensive history taking from the couple, gy- naecological examination and measurement of relevant hormo- nal levels in the women and complete andrological work-up of the male partner. The couples were counselled on the treatment procedure including relevant risks and possible complications. Thereafter medical pre-treatment was commenced with either a short/long agonist protocol or an antagonist protocol with hCG application for ovulation induction and oocyte puncture. After fertilisation with IVF or ICSI, in the conservative ART cycle (con- trol group) a maximum of 3 × 2PN stages was cultured for 2 – 3 days up until transfer, while in the DMW cycle a maximum of 6 × 2PN stages was selected for 2 – 5 day in vitro culture. Inclusion and exclusion criteria 400 patients between the ages of 21 and 45 years who received assisted reproduction (IVF or ICSI) at least twice between 2006 and 2013 were included in this retrospective cohort study and were allocated to either control (n = 200, CG) or experimental groups (n = 200, EG). A further inclusion requirement was that the observed treatment cycle was not discontinued. The two last treatments were taken into account (1st cycle = CG1, 2nd cycle = CG2 vs. 1st cycle = EG1 and 2nd cycle EG2 [EG2 hereafter denoted as DMW]). We excluded patients who were only treated accord- ing to the DMW, which lead to many patients not automatically entering the analysis especially between 2011 and 2013. Inclusion criteria for the EG were at least one standard protocol treatment (EG1) and one treatment according to the DMW (DMW). Inclusion criteria for the CG were at least two standard treatment cycles (CG1 and CG2). Through matching the groups in this way it was not only possible to compare CG vs. EG, but also to perform a patient-internal analysis by comparing previous with subsequent cycles in individual patients. As an example, in a patient in the EG group who was treated four times, twice with each protocol, only the last standard cycle and last DMW cycle were recorded and analysed with all parameters. In CG patients the last two standard cycles were captured. This

was chosen in order to avoid positive selection bias with- in the collective and to enable uniform analysis of patients with different numbers of treatment cycles. Parameters The following parameters were captured: Patient age, number of years trying to fall pregnant, body mass index (BMI), number of previous live births, abortions and evacuations, partner ʼs andro- logical results, method of insemination (IVF or ICSI), date of puncture and embryo transfer, and protocol type. 691 Kliebisch TK et al. The German Middleway … Geburtsh Frauenheilk 2016; 76: 690 –698 Original Article Additional parameters measured: " Base-line hormone levels between day 3 and 5 of the cycle: " Luteinising hormone (LH) " Follicle-stimulating hormone (FSH) " Anti-Müllerian hormone (AMH) " Thyroid stimulating hormone (TSH) " Testosterone " Prolactin " Dehydroepiandrosterone sulphate (DHEA‑S) " LH/FSH ratio (calculated) " On the day of human chorionic gonadotropin (hCG) application for ovulation induction: " Oestradiol (E2)a n d " Luteinising hormone (LH) Also documented and analysed: " Medication for stimulation: " Recombinant follicle-stimulating hormone (rFSH) " Recombinant luteinising hormone (rLH) " HMG " Recombinant human chorionic gonadotropin (rhCG) " Corifollitropin alfa " HCG " Gonadotropin-releasing hormone agonists (GnRH agonists) " Gonadotropin-releasing hormone antagonists (GnRH antag- onists) " Contraceptive in the previous cycle (if applicable) " Thyroid hormone substitution (if applicable) " Anticoagulants (if applicable) Further parameters analysed included: the number of punctured oocytes and their quality through the course of treatment (defec- tive oocytes, germinal vesicles [GV], 1st meiotic division attained [MI], mature oocytes [ “nachgereift” = N], 2nd meiotic division at- tained [MII] and following insemination [2PN]), the numbers of cryopreserved 2PNs, cultured 2PNs, planned and transferred em- bryos, and the quality of embryos (embryo grading). The first hCG value was obtained 13 days after oocyte puncture, the second a further 7 days later. Also captured were the first ultrasound re- sult, possible twin constellation with its peculiarities such as van- ishing twins, and lastly data on the birth (duration of pregnancy, sex, weight and abnormalities of the newborn, type of delivery). Parameters captured on history included: treatment outside of our centre, discontinued treatments in the past and reasons in each case (ovarian hyperstimulation syndrome [OHSS], no 2PNs); number of cycles with optimally timed sexual intercourse (“Geschlechtsverkehr zum Optimalen Zeitpunkt”, VZO), number of intrauterine inseminations (IUI), in vitro fertilisations (IVF), in- tracytoplasmic sperm injections (ICSI) and cryotransfers; demo- graphic factors such as marital status and nicotine consumption; gynaecological treatments such as laparoscopy (LSK), hysteros- copy (HSK), chromopertubation; gynaecological diseases such as endometriosis, uterine myomas, polycystic ovarian syndrome (PCO) and uterine position. Statistical analysis All calculations, statistical and descriptive analyses for this study were conducted using the IBM SPSS Statistics Version 22.0.0.0 and with support and specialist expertise from .05 Statistikbera- tung Düsseldorf. Statistical analysis was by means of the t-test and Levene ʼst e s t (comparison of means), as well as Pearson ʼs χ 2-test, the Mann- Whitney and Wilcoxon tests (group analyses) and ANOVA analy- ses (linear regression analyses)/the Omnibus test (logistic regres- sion analyses). Correlation analyses were also performed. Statis- tical significance was defined as p < 0.05.

! Statistical description of primary outcome measures Primary outcomes were pregnancy rate (pregnancy progressing with rising hCG and positive results on 1st ultrasound), birth rate (birth of one or more children) and absolute number of children born/live births. In the CG1 pregnancy was achieved in 18 % of cases (1st hCG x ̄ [average] = 152 mU/ml, 2nd hCG x̄ = 2373 mU/ml, n = 36), in EG1 in 16.5 % (1st hCG x ̄ = 86 mU/ml, 2nd hCG x ̄ = 1384 mU/ml, n = 33), in CG2 in 31 % (1st hCG x ̄ = 146 mU/ml, 2nd hCG x̄ = 2366 mU/ml, n = 62) and in the DMW group in 35 % of cases (1st hCG x̄ = 222 mU/ml, 2nd hCG x̄ = 2530 mU/ml, n = 70). In CG1 the birth rate of 5.5 % (11 of n = 200) was equal to the num- ber of children born, and in EG1 was 5 % (10 of n = 200). The birth rate in CG2 was 28.5 % (57 of n = 200) with 66 the absolute num- ber of children born. The birth rate in the DMW group was 30.5 % (61 of n = 200), 75 the absolute number of children born. The primary outcomes are shown in l " Fig. 1 a and b. Only pa- tients who achieved a progressing pregnancy in the first or sec- ond cycles were included in the analysis. Purely biochemical pregnancies were not included. Group similarity and unchanged parameters CG vs. EG These parameters refer to the end of the data collection period and are therefore not specific for individual patient treatment cycles. Only CG and EG are differentiated in order to reflect the similarity of the groups. It was found that the CG and EG were not significantly different with respect to total number of ART cycles (IVF and ICSI), VZO, IUIs, number of cryotransfers, smoking habit, occurrence of OHSS, absent 2PN cells, discontinued treatments, marital status, incidence of endometriosis, HSKs, chromopertubations, presence of uterine myomas, uterus position or history of PCO. There were however significant differences between the groups in the number of DMW cycles (EG = 1.44 ± 0.77, p = 0.000), since only the EG underwent DMW, and in the number of cryotransfers (CG = 1.05 ± 1.39, EG = 1.61 ± 1.82, p = 0.001) and history of LSKs (CG = 30.5 %, EG = 40 %, p = 0.047). Group similarity and comparison of first cycle CG1 vs. EG1 For the first cycle both groups were treated according to the stan- dard protocol and here too the similarity of the groups was shown. The groups did not differ significantly with respect to: age; BMI; E 2 and LH levels on hCG application; use of rFSH, rFSH + rLH, GnRH agonists and antagonists; laboratory parameters (LH/FSH, AMH, TSH, testosterone, LH, DHEA); number of oocytes MI; cultivated 2PNs; transferred and cultivated embryos; 2nd hCG; day between puncture and transfer, and days from transfer to birth; quality of the first, second and third embryos; 1st ultra- sound result, pregnancy and birth rates or number of live births. The groups did differ significantly for: dose of FSH (CG = 2509 ± 1357 IU, EG = 2070 ± 1058 IU, p = 0.0003), prolactin level (CG = 14.52 ± 7.34, EG = 16.18 ± 7.54, p = 0.031), FSH level (CG = 8.32 ± 4.14, EG = 7.24 ± 3.48, p = 0.006), number of punc- tured oocytes (CG = 7.75 ± 3.82, EG = 9.12 ± 4.05, p = 0.001), MII oocytes (CG = 6.02 ± 3.38, EG = 7.21 ± 3.60, p = 0.001), 2PNs 692 Kliebisch TK et al. The German Middleway … Geburtsh Frauenheilk 2016; 76: 690 –698 GebFra Science (CG = 3.77 ± 2.47, EG = 4.88 ± 2.88, p = 0.001), cryopreserved 2PNs (CG = 1.40 ± 2.20, EG = 2.46 ± 2.91, p = 0.000), 1st hCG (CG = 151.58 ± 181.95, EG = 83.43 ± 59.86, p = 0.046) and type of ART (CG = 88.5 % ICSI, EG = 80 % ICSI, p = 0.020). Comparison of groups and second cycle CG2 vs. EG2 (DMW) There were no significant differences in the primary outcomes, although higher pregnancy rates (+ 4 %), birth rates (+ 2 %) and more live births (+ 9) were achieved in the DMW despite fewer embryos transferred (CG = 1.96 ± 0.62, EG = 1.83 ± 0.38, p = 0.009) (l " Table 1). Group comparison and 1st and 2nd cycles EG1 vs. EG2 (DMW) in individual patients Since this comparison considered different cycles within the same individuals, one standard (EG1) and one DMW (EG2) proto- col, it was possible to show that patients benefitted significantly from the new method despite their inevitably significantly in- creased age (EG1 = 35.43 ± 3.99, EG2 = 36.45 ± 3.86, p = 0.000). Significantly higher pregnancy rates (+ 18.5 %) and birth rates (+ 25.5 %) were achieved despite significantly fewer embryos being transferred (EG1 = 1.97 ± 0.46, EG2 = 1.83 ± 0.38, p = 0.000) (l " Table 2). Regression analyses of EG1, CG2 and EG2 (DMW) Effect of embryo quality on pregnancy occurrence A multiple logistic regression was also performed with the qual- ity (A to C) of the first, second and third transferred embryo as predictors of pregnancy occurrence. For EG1 the overall variance of the logistic regression was χ 2 (6) = 19.9, p < 0.01. The quality of the second embryo was a sig- nificant predictor (p < 0.05). For CG2 the overall variance of the logistic regression was χ 2 (6) = 28.13, p < 0.01 with the quality of the first and second em- bryos being significant (p < 0.01). No significant correlation was found for EG2. Effect of ultrasound results on occurrence of birth In EG1 (χ2[3] = 106.77, p < 0.01), CG2 (χ2[3] = 190.67, p < 0.01) and EG2 (χ2[3] = 178.77, p < 0.01) the first ultrasound result was a sig- nificant predictor of a birth. Regression analysis EG2 (DMW) Effect of cultivated 2PN stages and patient age on number of blastocysts generated The overall variance of the multiple linear regression was R2 = 0.25 (F = 32.67, p < 0.01). The number of cultured 2PN cells (β = 0.49, p < 0.01) was a significant predictor. Effect of number of cultivated 2PN stages on the first ultrasound result at 7 weeks gestation Overall, in CG2 23 % achieved intact singleton pregnancies. 2.5 % occurred after culture of 1, 14 % after culture of 2, and 6.5 % on culture of 3 × 2PNs. 6.5 % had intact twin pregnancies with 6 % oc- curring after culture of 2, and 0.5 % on culture of 3 × 2PN stages. In EG2 24.5 % of cases had a singleton pregnancy with 1 % occur- ring after culture of 1, 2 % after 2, 2.5 % after 3, 7.5 % after 4, 7 % after 5 and 4.5 % after culture of 6 2PNs. Twin pregnancies (8 %) occurred in 0.5 % after culture of 2, 6.5 % after 4 and 0.5 % after cul- ture of both 5 and 6 × 2PNs respectively ( l " Table 3).

! The introduction of the DMW is a further step towards increasing the pregnancy rate with ART despite restrictive German legisla- tion. We have shown that these patients can benefit significantly from new techniques: in comparison to previous treatment cycles, significantly higher pregnancy and birth rates were at- tained despite patents being significantly older and receiving sig- nificantly fewer transferred embryos. It was clearly demonstrat- ed that ultimately the number of generated 2PN stages and the quality of transferred embryos were decisive. Due to more embryos being cultivated, the number of transferred embryos could be reduced while achieving higher pregnancy and Pregnancy Birth 18 11 10 66 75 5.5 16.5 5 31 28.5 35 30.5 Percent Number o f children born 40 30 20 10 0 80 60 40 20 0 CG1ab CG1EG1 EG1 CG2 CG2 DMW DMW Fig. 1 a and ba Pregnancy and birth rates within the study collectives. Representation of the pregnancy and birth rates of the control group (CG1 and 2, the first and last conservative cycles respectively) and the experimen- tal group (EG1 initial conservative schema, then DMW). b Number of live births in the treatment groups. The greatest number of children were born following treatment according to the DMW schema fol- lowed by conservative treatment. 693 Kliebisch TK et al. The German Middleway … Geburtsh Frauenheilk 2016; 76: 690 –698 Original Article birth rates. In this context it was the embryo quality and not the day of transfer that was decisive. Another study has also shown that embryo quality is the most significant factor for the cumula- tive live birth rate in the context of eSET, although here the odds ratio in the presence of ≥ 3 “top” embryos was 2.66 [8]. Better trophectoderm morphology, low patient age and the best possible blastocyst developmental stage had the greatest influ- ence on clinical pregnancy and live birth rates in an American study. Thus primarily the trophectoderm morphology and blas- tocyst stage should be used to select the best embryo for transfer. Blastocyst stage was associated with live birth rates of 50 % (hatching), 49.5 % (expanded) and 36.7 % (early) [9]. A further study also showed that age and number of “top” em- bryos are significantly associated with occurrence of clinical pregnancy after eSET. The authors recommend performing eSET preferably on day 3 [10]. Practice algorithms should be developed to predict success with the DMW. The number of cultivated 2PNs had a stronger influ- ence on blastocyst extraction than age, hence it is older patients in particular who benefit from the DMW when enough oocytes can be harvested and cultivated. It was shown that culture of 4 × 2PN stages in the context of DMW resulted in the highest pregnancy rates; overall 14 % of the Table 1 Demographics and descriptive statistics by number of patients (n) for DMW group vs. conservative treatment (CG2). Representation of mean values ± standard deviations or percentages; in brackets absolute incidences with p < 0.05. CG2 (n = 200) DMW (n = 200) p-value Age 36.94 ± 4.03 (200) 36.45 ± 3.86 (200) BMI 23.39 ± 4.01 (199) 22.88 ± 3.91 (200) E2 at hCG application in pg/ml 1 827 ± 1 358 (200) 1 702 ± 1 124 (200) LH at hCG application in µIU/ml 3.90 ± 2.84 (200) 3.34 ± 3.18 (200) Total FSH dose in IU 2 505 ± 1 419 (200) 1 987 ± 1 158 (200) < 0.05 GnRH antagonist in mg 0.52 ± 0.76 (70) 0.77 ± 0.81 (103) < 0.05 GnRH agonist in mg 4.93 ± 4.09 (128) 3.30 ± 3.70 (97) LH/FSH ratio in µIU/ml 0.87 ± 0.44 (193) 0.90 ± 0.38 (191) AMH in µg/l 2.00 ± 2.76 (124) 1.99 ± 1.98 (169) TSH in µIU/ml 1.72 ± 0.82 (197) 1.92 ± 1.87 (193) Testosterone in ng/ml 0.27 ± 0.18 (195) 0.25 ± 0.16 (190) Prolactin in ng/ml 15.14 ± 12.09(190) 15.83 ± 7.57 (190) FSH in µIU/ml 8.32 ± 4.02 (193) 7.04 ± 2.32 (192) < 0.05 LH in µIU/ml 6.27 ± 2.54 (196) 6.05 ± 2.51 (194) DHEA in µg/l 141.24 ± 69.87 (191) 151.34 ± 70.85 (192) Number of oocytes 7.6 ± 4.36 (200) 9.89 ± 3.96 (200) < 0.05 MI oocytes 0.87 ± 1.16 (104) 0.99 ± 1.33 (102) MII oocytes 5.84 ± 3.59 (200) 7.88 ± 3.50 (200) < 0.05 2PNs 3.82 ± 2.68 (200) 5.36 ± 2.90 (200) < 0.05 2PNs planned for culture 2.22 ± 0.49 (200) 5.18 ± 0.95 (200) < 0.05 Cultivated 2PNs 2.01 ± 0.63 (200) 3.96 ± 1.40 (200) < 0.05 Transferred embryos 1.96 ± 0.62 (200) 1.83 ± 0.38 (200) < 0.05 1st hCG in mU/ml 145.78 ± 122.59 (62) 221.52 ± 458.98 (62) 2nd hCG in mU/ml 2 365.75 ± 1 740.55 (62) 2 529.6 ± 1 915.91 (62) Difference in days between puncture and transfer 2.31 ± 0.46 (200) 3.93 ± 1.18 (200) < 0.05 Days between transfer and birth (duration of pregnancy) 254.95 ± 15.39 (57) 245.56 ± 23.20 (61) < 0.05 Quality of 1st embryo (200) (198) < 0.05 " A 48 % (96) 67.5 % (135) " B 43 % (86) 22 % (44) " C 9 % (18) 9.5 % (19) Quality of 2nd embryo (156) (161) " A 32.5 % (64) 43,5 % (87) " B 34.5 % (69) 21 % (42) " C 11.5 % (23) 16 % (32) Quality of 3rd embryo (33) (0) < 0.05 " A 24 % (8) " B 55 % (18) " C 21 % (7) Blastocysts (A) and morula stages (C) are assessed in quality criteria ART type (ICSI) 92 % (184) 81.5 % (163) < 0.05 " 1st US (62) (70) " Singleton 74 % (46) 71 % (50) " Twin 21 % (13) 23 % (16) " Abortion 5 % (3) 6 % (4) Pregnancy 31 % (62) 35 % (70) Birth 28.5 % (57) 30.5 % (61) Number of live births 66 75 694 Kliebisch TK et al. The German Middleway … Geburtsh Frauenheilk 2016; 76: 690 –698 GebFra Science 32.5 % pregnancies achieved were due to the culture of 4 × 2PN stages. 12.5 % resulted from culture of > 4 × 2PN stages. On the one hand, the fact that culture of 4 × 2PN stages achieved the highest pregnancy rates can be explained by the patients hav- ing a better prognosis. On the other hand, this possible statistical bias is put into perspective by the fact that patients with worse prognosis were also included in the group when the planned number of 6 × 2PNs to be cultured was not achieved. Naturally, it must always be taken into consideration that surplus embryos will be produced in the process of culturing 2PN stages according to the DMW, and these must then be frozen. Since this rate of sur- plus embryos should be kept as low as possible, decisions should be made on an individual basis after careful history taking and taking previous stimulations and transfers into consideration. A number of blood values and medications were identified as predictors: it was statistically advantageous, in terms of generat- ing a high number of oocytes, to achieve the highest possible E 2 level together with an initially high AMH level and the lowest possible LH. High doses of urofollitropin + lutropin and GnRH antagonists with resultant long stimulation were negative predictors of oo- cyte number. This patient collective demonstrated low AMH lev- els and a poor ovarian response to stimulation, which explained the long duration of stimulation using the antagonist protocol. Table 2 Demographics and descriptive statistics by number of patients (n) for patients undergoing the DMW method in their second cycle. Representation of mean values ± standard deviations or percentages; in brackets absolute incidences with p < 0.05. EG1 DMW p-value Age 35.43 ± 3.99 (200) 36.45 ± 3.86 (200) < 0.05 BMI 22.83 ± 3.98 (200) 22.88 ± 3.91 (200) E2 at hCG application in pg/ml 1 976 ± 1 292 (197) 1 702 ± 1 124 (200) < 0.05 LH at hCG application in µIU/ml 3.91 ± 3.50 (197) 3.34 ± 3.18 (200) Total FSH dose in IU 2 070 ± 1 058 (200) 1 987 ± 1 158 (200) GnRH antagonist 0.44 ± 0.70 (69) 0.77 ± 0.81 (103) < 0.05 GnRH agonist 5.28 ± 3.93 (141) 3.30 ± 3.70 (97) < 0.05 LH/FSH ratio in µIU/ml 0.93 ± 0.48 (192) 0.90 ± 0.38 (191) AMH in µg/l 2.20 ± 1.98 (192) 1.99 ± 1.98 (169) < 0.05 TSH in µIU/ml 1.87 ± 0.98 (193) 1.92 ± 1.87 (193) Testosterone in ng/ml 0.26 ± 0.17 (190) 0.25 ± 0.16 (190) < 0.05 Prolactin in ng/ml 16.18 ± 7.54 (190) 15.83 ± 7.57 (190) FSH in µIU/ml 7.24 ± 3.48 (192) 7.04 ± 2.32 (192) LH in µIU/ml 6.41 ± 3.97 (194) 6.05 ± 2.51 (194) DHEA‑S in µg/l 152.07 ± 71.79 (191) 151.34 ± 70.85 (192) Number of oocytes 9.12 ± 4.05 (200) 9.89 ± 3.96 (200) < 0.05 MI oocytes 0.85 ± 1.17 (98) 0.99 ± 1.33 (102) MII oocytes 7.21 ± 3.60 (197) 7.88 ± 3.50 (200) < 0.05 2PNs planned for culture 2.11 ± 0.37 (200) 5.18 ± 0.95 (200) < 0.05 Cryopreserved 2PNs 2.46 ± 2.91 (108) 1.06 ± 2.08 (53) < 0.05 Cultivated 2PNs 2.03 ± 0.49 (199) 3.96 ± 1.40 (200) < 0.05 Transferred embryos 1.97 ± 0.46 (200) 1.83 ± 0.38 (200) < 0.05 1st hCG in mU/ml 83.43 ± 59.86 (36) 221.52 ± 458.98 (62) < 0.05 2nd hCG in mU/ml 1 513.13 ± 1 327.49 (36) 2 529.6 ± 1 915.91 (62) < 0.05 Difference in days between puncture and transfer 2.35 ± 0.48 (200) 3.93 ± 1.18 (200) < 0.05 Days between transfer and birth (duration of pregnancy) 252,9 ± 12,78 (10) 245.56 ± 23.20 (61) Quality of 1st embryo (200) (198) < 0.05 " A 52 % (104) 67.5 % (135) " B 39.5 % (79) 22 % (44) " C 8.5 % (17) 9.5 % (19) Quality of 2nd embryo (175) (161) " A 27 % (54) 43.5 % (87) " B 46 % (92) 21 % (42) " C 12.5 % (25) 16 % (32) Quality of 3rd embryo (20) (0) < 0.05 " A 25 % (5) " B 50 % (10) " C 25 % (5) Blastocysts (A) and morula stages (C) are assessed in quality criteria ART type (ICSI) 80 % (160) 81.5 % (163) < 0.05 " 1st US (33) (70) " Singleton 42 % (14) 71 % (50) " Twin 12 % (4) 23 % (16) " Abortion 46 % (15) 6 % (4) Pregnancy 16.5 % (33) 35 % (70) < 0.05 Birth 5 % (10) 30.5 % (61) < 0.05 Number of live births 10 75 < 0.05 695 Kliebisch TK et al. The German Middleway … Geburtsh Frauenheilk 2016; 76: 690 –698 Original Article Other study groups have also demonstrated that a long agonist protocol, which constituted the majority of stimulations up until 2012, is better suited to the production of a greater number of mature follicles in the context of controlled ovarian stimulation [11]. The available data on FSH dose used clearly reflect the shift from agonist to antagonist protocols [12]. It has also been shown that higher implantation rates (OR 1.36) and a higher live birth rate (OR 1.33) are achieved with GnRH agonists in the context of eSET [13]. The agonist protocol carries the clinical risk of rapidly progressing ovarian hyperstimulation syndrome (OHSS), since hCG must be given for final oocyte maturation. The antagonist protocol is associated with a pronounced reduction in OHSS inci- dence through its use of physiological LH, which is triggered by the application of a GnRH agonist [14, 15]. Apart from the abso- lute number of mature follicles, further important clinical param- eters on comparison of agonist and antagonist protocols include: clinical and progressing pregnancy, abortion rate and live birth rate. To date, for these parameters no significant differences be- tween the protocols have been demonstrated [15]. At the same time it should not be forgotten that not every patient will automatically benefit from blastocyst transfer. While a small significant advantage for live birth rate has been shown with blastocyst transfer on day 5 to 6, a higher cumulative pregnancy rate was achieved with transfers on day 2 to 3 [16]. The authors of this study felt the result could be explained firstly by the fact that more cryotransfers took place in the classical ART cycles, and sec- ondly because the rate of treatment discontinuations, i.e. transfer refusals, was lower [16]. eSET should be used with circumspec- tion particularly in patients with limited ovarian reserves, i.e. poorer prognosis, since in some cases worse results are to be ex- pected [17]. In 2012 the American Society for Reproductive Medicine (ASRM) drew the following conclusions with respect to eSET: eSET should chiefly be offered to patients with a good prognosis/recipients of donor oocytes. IVF centres should only promote eSET when pa- tients are appropriately informed. Also, developments in embryo selection, e.g. through time-lapse techniques, should increase the use of eSET [18]. It is thus all the more important that decision-making regarding the use of DMW for a particular patient is well-informed and in- dividualised. Data acquisition in our study was retrospective and from a single centre so that applicability to other centres is limited. From a methodological perspective it was advantageous to compare pre- vious and subsequent treatment cycles within individual pa- tients, since this provided an internal control. Nevertheless this

also affected the results through only including patients who received at least one further treatment cycle. Patients who were treated by the DMW in their first cycle and who immediate- ly fell pregnant were thus excluded. This explains the extremely low pregnancy rate of initial treatment cycles, both in the CG and EG, which accordingly was lower than the usual averages of the DIR and the study centre: patients included were almost exclu- sively those who did not fall pregnant in the first cycle or who had abortions. In addition the selection process resulted in the analysis espe- cially of older patients (CG2 = 37, EG2/DMW = 36.5 years) with a relatively higher number of cycles (CG = 3.76, EG = 3.75 ART cycles per patient) and longer-standing wish for children (CG2 = 3.65 ± 2.28, EG2/DMW = 4.96 ± 2.36). In some cases pre- vious fertility treatments had been successful. 28.5 % of women in the CG2 and 22.5 % in the DMW group already had children. 16 % of these in the CG2 and 12.5 % in the DMW group were the

of successful fertility treatment. If one considers that our study population was one that took rel- atively long to fall pregnant and had a strongly negative selection, then the pregnancy and birth rates in the DMW group are very Table 3 Cross tabulation. Representation of the first ultrasound result at 7 weeks gestation with respect to the number of cultivated 2PNs in CG2 and DMW. Number of cultivated 2PNs (2nd cycle) × 1st ultrasound (2nd cycle) CG2 and DMW 1st ultrasound (2nd cycle) Total No ultrasound Intact, singleton Intact, twins Abortion CG2 Number of cultivated 2PNs 1 Count 34 5 0 0 39 % of Total 17.0 % 2.5 % 0.0 % 0.0 % 19.5 % 2 Count 78 28 12 3 121 % of Total 39.0 % 14.0 % 6.0 % 1.5 % 60.5 % 3 Count 26 13 1 0 40 % of Total 13.0 % 6.5 % 0.5 % 0.0 % 20.0 % Total Count 138 46 13 3 200 % of Total 69.0 % 23.0 % 6.5 % 1.5 % 100.0 % DMW Number of cultivated 2PNs 1 Count 7 2 0 0 9 % of Total 3.5 % 1.0 % 0.0 % 0.0 % 4.5 % 2 Count 24 4 1 0 29 % of Total 12.0 % 2.0 % 0.5 % 0.0 % 14.5 % 3 Count 20 5 0 0 25 % of Total 10.0 % 2.5 % 0.0 % 0.0 % 12.5 % 4 Count 39 15 13 2 69 % of Total 19.5 % 7.5 % 6.5 % 1.0 % 34.5 % 5 Count 19 14 1 0 34 % of Total 9.5 % 7.0 % 0.5 % 0.0 % 17.0 % 6 Count 24 9 1 0 34 % of Total 12.0 % 4.5 % 0.5 % 0.0 % 17.0 % Total Count 133 49 16 2 200 % of Total 66.5 % 24.5 % 8.0 % 1.0 % 100.0 % 696 Kliebisch TK et al. The German Middleway … Geburtsh Frauenheilk 2016; 76: 690 –698 GebFra Science promising indeed, supporting the idea of greater success with a change to this procedure. The results demonstrate that individual patients benefited from the procedure. With 61 vs. 70 actual pregnancies in the two groups respectively (pregnancy rate 31 vs. 35 %) statistical power was not strong enough to show a significant benefit for EG over CG. Ultimately the study only shows trends favouring the DMW, which need to be confirmed in larger patient collectives or anal- yses of individual cycles. Since in both groups we only analysed cycles involving an em- bryo transfer, our results are highly consistent in terms of the tar- get parameters (embryo quality, pregnancy rate etc.). Our collective was too small to demonstrate a reduced rate of multiple pregnancy with DMW. In 2006 a Dutch study suggested that eSET be used in all patients if possible in order to avoid multiple pregnancy; here, however, the pregnancy rate was concurrently halved [19]. Since then it has been shown that particularly patients older than 35 who are treated with eSET and extended embryo selection, i.e. with genetic screening before embryo transfer, have significantly fewer multiple pregnancies (6.8 vs. 21 %) while also having a sig- nificantly higher live birth rate per embryo transfer (17 vs. 10.6 %). For this study it was however necessary to implement re- laxed policies on embryo selection, optimum culture systems, cryopreservation and aneuploidy screening of blastocysts with analysis of all 24 chromosomes [20]. It has also since been demonstrated that, unlike Double Embryo Transfer (DET), multiple pregnancy can be significantly reduced (in one study to 35 %) by the use of eSET [21]. Other study groups have also managed to show the value of eSET by demonstrating risk reduction whilst maintaining IVF success rates [22]. However, what should the approach be when a patient insists on transfer of multiple embryos despite the risk of multiple preg- nancy in order to increase her chances of pregnancy within a sin- gle treatment attempt, or when the patient makes a conscious decision in favour of having twins? Patients should be empow- ered to make an autonomous decision through the concept of shared decision making, including comprehensive counselling on the risks of multiple pregnancy and the associated increased financial burden [23]. A shift in patient preference for twin pregnancy during the course of treatment has been described, whereby preferences usually adapt to the pregnancy situation: Whereas patients who are already pregnant tend to prefer what they are expecting (sin- gleton or twins), before embryo transfer a greater percentage of patients show a preference for twins in order to fulfil their wish for children as soon as possible [24]. This illustrates the impor- tance of comprehensive counselling. In summary, it can be stated that the DMW increases the proba- bility of pregnancy. The original intention to improve pregnancy rates while reducing the risk of twins – and associated health, personal, and financial implications – can only be realised when the number of embryos transferred is reduced. Northern Europe, Australia and Japan are increasingly striving for the permanent establishment of eSET to reduce high rates of twins while maintaining high pregnancy rates in general. Coun- tries such as Sweden, Finland and Belgium have demonstrated that twin pregnancies can be reduced to less than 10 % compared to otherwise generally unchanged rates of ~ 20 % in many Euro- pean countries [25, 26]. In Sweden reproductive health practi- tioners are only allowed to transfer single embryos, except in ex- traordinary circumstances. In Belgium the first cycle has to con- stitute an eSET in patients less than 36 years of age. In Finland there are not yet any statutory provisions [27]. In Germany the rule forbidding routine, intentional stockpiling of embryos remains unchanged. The implementation of the DMW would offer German reproduc- tive medicine practitioners the possibility of achieving higher pregnancy rates combined with reduced risk of twins, however this can only be realised if it is accepted that viable surplus em- bryos be frozen and stored. After thorough, individualised pa- tient counselling and consent, taking the experiences of previous ART cycles into account, the initial intent is the transfer of one viable embryo. Since the quality of mature oocytes produced dur- ing stimulation can vary from cycle to cycle, patients must be fully informed about the need to freeze-store viable embryos in the event of more embryos being cultured than are necessary for single embryo transfer. On this point too, the 1990 ESchG is urgently in need of reform. Only the national legislature is in a position to regulate these fun- damental questions around the beginnings of human life; thus a new reproductive medicine act that is up-to-date with current medical and scientific research – however, which should not be seen as a mediator for future unrestricted embryo culture and stockpiling – is urgently required.

! The DMW offers the possibility of increasing pregnancy rates particularly in patients with adequate ovarian reserves. It should be considered when adequate numbers of oocytes can be gener- ated, since these patients benefit significantly from the transfer of fewer embryos of the highest possible quality.

! We thank .05 Statistikberatung – Diplom-Psychologen Hoffmann Ostapczuk Ullrich Partnerschaftsgesellschaft, Düsseldorf, for the statistical analysis and support. Conflict of Interest ! The authors declare that no conflict of interest exists.

1 Snick HK, Snick TS, Evers JL et al. The spontaneous pregnancy prognosis in untreated subfertile couples: the Walcheren primary care study. Hum Reprod 1997; 12: 1582 – 1588 2 Diedrich K, Strowitzki T, Kentenich H. Stand der Reproduktionsmedizin in Deutschland. Geburtsh Frauenheilk 2010; 70: 355 – 360 3 Bundesministerium der Justiz. Gesetz zum Schutz von Embryonen (Em- bryonenschutzgesetz – ESchG). Embryonenschutzgesetz vom 13. De- zember 1990 (BGBl. I S. 2746), das zuletzt durch Artikel 1 des Gesetzes vom 21. November 2011 (BGBl. I S. 2228) geändert worden ist 4 Bals-Pratsch M, Frommel M. Wandel in der Implementation des Deut- schen Embryonenschutzgesetzes. J Reproduktionsmed Endokrinol 2010; 7: 87 – 95 5 Frommel M. Deutscher Mittelweg in der Anwendung des Embryonen- schutzgesetzes (ESchG) mit einer an den aktuellen wissenschaftlichen Kenntnisstand orientierten Auslegung der für die Reproduktionsmedi- zin zentralen Vorschrift des § 1, Abs. 1, Nr. 5 ESchG. J Reproduk- tionsmed Endokrinol 2007; 4: 27 – 33 697 Kliebisch TK et al. The German Middleway … Geburtsh Frauenheilk 2016; 76: 690 –698 Original Article 6 Cupisti S, Muller A, Hildebrandt T et al. Culture of individually required number of 2-pronuclei-stage oocytes – patient participation in deci- sion-making is in accordance with the aim of avoiding surplus embryo freezing. Geburtsh Frauenheilk 2014; 74: 157 – 160 7 Das Jahrbuch des Deutschen IVF-Registers 2013. J Reproduktionsmed Endokrinol 2014; 11: 6 – 51 8 Niinimaki M, Veleva Z, Martikainen H. Embryo quality is the main fac- tor affecting cumulative live birth rate after elective single embryo transfer in fresh stimulation cycles. Eur J Obstet Gynecol Reprod Biol 2015; 194: 131 – 135 9 Thompson SM, Onwubalili N, Brown K et al. Blastocyst expansion score and trophectoderm morphology strongly predict successful clinical pregnancy and live birth following elective single embryo blastocyst transfer (eSET): a national study. J Assist Reprod Genet 2013; 30: 1577– 1581 10 Sifer C, Sermondade N, Poncelet C et al. Biological predictive criteria for clinical pregnancy after elective single embryo transfer. Fertil Steril 2011; 95: 427 – 430 11 Shrestha D, La X, Feng HL. Comparison of different stimulation protocols used in in vitro fertilization: a review. Ann Transl Med 2015; 3: 137 12 Das Jahrbuch des Deutschen IVF-Registers 2011. J Reproduktionsmed Endokrinol 2012; 9: 453 – 484 13 Grow D, Kawwass JF, Kulkarni AD et al. GnRH agonist and GnRH antag- onist protocols: comparison of outcomes among good-prognosis pa- tients using national surveillance data. Reprod Biomed Online 2014; 29: 299– 304 14 Devroey P, Polyzos NP, Blockeel C. An OHSS-free clinic by segmentation of IVF treatment. Hum Reprod 2011; 26: 2593 – 2597 15 Al-Inany HG, Yousseff MA, Aboulghar M et al. Gonadotrophin-releasing hormone antagonists for assisted reproductive technology. Cochrane Database Syst Rev 2011; 5: CD001750 16 Glujovsky D, Blake D, Farquhar C et al. Cleavage stage versus blastocyst stage embryo transfer in assisted reproductive technology. Cochrane Database Syst Rev 2012; 7: CD002118 17 Gleicher N, Kushnir VA, Barad DH. Is it time for a paradigm shift in understanding embryo selection? Reprod Biol Endocrinol 2015; 13: 3 18 Practice Committee of Society for Assisted Reproductive Technology; Practice Committee of American Society for Reproductive Medicine. Elec- tive single-embryo transfer. Fertil Steril 2012; 97: 835 – 842 19 van Montfoort AP, Fiddelers AA, Janssen JM et al. In unselected patients, elective single embryo transfer prevents all multiples, but results in significantly lower pregnancy rates compared with double embryo transfer: a randomized controlled trial. Hum Reprod 2006; 21: 338 – 343 20 Ubaldi FM, Capalbo A, Colamaria S et al. Reduction of multiple pregnan- cies in the advanced maternal age population after implementation of an elective single embryo transfer policy coupled with enhanced em- bryo selection: pre- and post-intervention study. Hum Reprod 2015; 30: 2097– 2106 21 Prados N, Quiroga R, Caligara C et al. Elective single versus double em- bryo transfer: live birth outcome and patient acceptance in a prospec- tive randomised trial. Reprod Fertil Dev 2015; 27: 794 – 800 22 Ismail L, Mittal M, Kalu E. IVF twins: buy one get one free? J Fam Plann Reprod Health Care 2012; 38: 252 – 257 23 Ezugwu E, der Burg SV. Debating elective single embryo transfer after in vitro fertilization: a plea for a context-sensitive approach. Ann Med Health Sci Res 2015; 5: 1 – 7 24 Fiddelers AA, Nieman FH, Dumoulin JC et al. During IVF treatment pa- tient preference shifts from singletons towards twins but only a few patients show an actual reversal of preference. Hum Reprod 2011; 26: 2092– 2100 25 Kupka MS, Ferraretti AP, de Mouzon J et al.; European IVF-Monitoring Consortium, for the European Society of Human Reproduction and Em- bryology. Assisted reproductive technology in Europe, 2010: results generated from European registers by ESHRE †. Hum Reprod 2014; 29: 2099– 2113 26 European IVF-Monitoring Consortium (EIM); European Society of Hu- man Reproduction and Embryology (ESHRE), Kupka MS, D ʼHooghe T et al. Assisted reproductive technology in Europe, 2011: results gener- ated from European registers by ESHRE. Hum Reprod 2016; 31: 233 – 248 27 Harbottle S, Hughes C, Cutting R et al.; Association of Clinical Embryolo- gists & The (ACE) British Fertility Society (BFS). Elective single embryo transfer: an update to UK Best Practice Guidelines. Hum Fertil (Camb) 2015; 18: 165 – 183 698 Kliebisch TK et al. The German Middleway … Geburtsh Frauenheilk 2016; 76: 690 –698 GebFra Science