STUDY QUESTION Is there an ideal imaging modality for the detection of uterosacral ligaments/torus uterinus (USL), rectovaginal septum (RVS) and vaginal deep endometriosis (DE) in women with a clinical history of endometriosis? SUMMARY ANSWER The sensitivity for the detection of USL, RVS and vaginal DE using MRI seems to be better than transvaginal ultrasonography (TVS), whilst the specificity of both were excellent. WHAT IS KNOWN ALREADY The surgical management of women with DE can be complex and requires advanced laparoscopic skills with maximal cytoreduction being vital at the first procedure to provide the greatest symptomatic benefit. Owing to a correlation of TVS findings with surgical findings, preoperative imaging has been used to adequately consent women and plan the appropriate surgery. However, until publication of the consensus statement by the International Deep Endometriosis Analysis Group in 2016, there were significant variations within the terms and definitions used to describe DE in the pelvis. STUDY DESIGN, SIZE, DURATION A systematic review and meta-analysis was conducted using Embase, Google Scholar, Medline, PubMed and Scopus to identify studies published from inception to May 2020, of which only those from 2010 were included owing to the increased proficiency of the sonographers and advancements in technology. PARTICIPANTS/MATERIALS, SETTING, METHODS All prospective studies that preoperatively assessed any imaging modality for the detection of DE in the USL, RVS and vagina and correlated with the reference standard of surgical data were considered eligible. Study eligibility was restricted to those including a minimum of 10 unaffected and 10 affected participants. MAIN RESULTS AND THE ROLE OF CHANCE There were 1977 references identified from which 10 studies (n = 1188) were included in the final analysis. For the detection of USL DE, the overall pooled sensitivity and specificity for all TVS techniques were 60% (95% CI 32–82%) and 95% (95% CI 90–98%), respectively, and for all MRI techniques were 81% (95% CI 66–90%) and 83% (95% CI 62–94%), respectively. For the detection of RVS DE, the overall pooled sensitivity and specificity for all TVS techniques were 57% (95% CI 30–80%) and 100% (95% CI 92–100%), respectively. For the detection of vaginal DE, the overall pooled sensitivity and specificity for all TVS techniques were 52% (95% CI 29–74%) and 98% (95% CI 95–99%), respectively, and for all MRI techniques were 64% (95% CI 40–83%) and 98% (96% CI 93–99%). Pooled analyses were not possible for other imaging modalities. LIMITATIONS, REASONS FOR CAUTION There was a low quality of evidence given the high risk of bias and heterogeneity in the included studies. There are also potential biases secondary to the risk of misdiagnosis at surgery owing to a lack of either histopathological findings or expertise, coupled with the surgeons not being blinded. Furthermore, the varying surgical experience and the lack of clarity regarding complete surgical clearance, thereby also contributing to the lack of histopathology, could also explain the wide range of pre-test probability of disease. WIDER IMPLICATIONS OF THE FINDINGS MRI outperformed TVS for the per-operative diagnosis of USL, RVS and vaginal DE with higher sensitivities, although the specificities for both were excellent. There were improved results with other imaging modalities, such as rectal endoscopy-sonography, as well as the addition of bowel preparation or ultrasound gel to either TVS or MRI, although these are based on individual studies. STUDY FUNDING/COMPETING INTEREST(S) No funding was received for this study. M.L. reports personal fees from GE Healthcare, grants from the Australian Women’s and Children’s Foundation, outside the submitted work. B.W.M. reports grants from NHMRC, outside the submitted work. G.C. reports personal fees from GE Healthcare, outside the submitted work; and is on the Endometriosis Advisory Board for Roche Diagnostics. REGISTRATION NUMBER Prospective registration with PROSPERO (CRD42017059872) was obtained.

Study question In couples with unexplained infertility, does IVF increase cumulative live birth rate and reduce multiple pregnancy rate compared to intrauterine insemination with ovarian stimulation (IUI-OS)? Summary answer There were no significant differences in cumulative live birth and multiple pregnancy rates between IVF and IUI-OS in couples with unexplained infertility. What is known already IVF and IUI-OS are widely used in managing unexplained infertility, especially in couples with a poor prognosis for natural conception. Although several randomized controlled trials (RCTs) have compared IVF versus IUI-OS, it remains inconclusive regarding which approach is more effective. Some of the RCTs did not define a time limit for follow up and others made the comparison on a per-cycle basis which provides a biased estimate in favour of IVF. Study design, size, duration We performed an individual participant data meta-analysis (IPD-MA) that synthesised available individual-level data comparing IVF and IUI-OS in unexplained infertility. We searched MEDLINE, EMBASE, CENTRAL, PsycINFO, CINAHL, and the Cochrane Gynaecology and Fertility Group Specialised Register of RCTs and included eligible RCTs that completed data collection before June 2021. We invited author groups of eligible studies to join the IPD-MA and share the deidentified IPD of their RCTs. Participants/materials, setting, methods RCTs that compared IVF/ICSI to IUI-OS in couples with unexplained infertility were included. The primary effectiveness outcome was cumulative live birth, defined by time to pregnancy leading to live birth. The primary safety outcome was the number of multiple pregnancies per participant. IPD were checked and standardised before synthesis. The quality of evidence was assessed using the Risk of Bias 2 tool. The analysis followed the intention-to-treat principle, and a two-stage IPD meta-analysis was performed. Main results and the role of chance Of eight potentially eligible RCTs, four shared individual-level data of 933 couples, of which 550 couples were allocated to IVF and 383 couples to IUI-OS. Two RCTs had a low risk of bias, one had some concerns, and one had a high risk of bias. Considering the time to pregnancy leading to live birth, the cumulative live birth rate was not significantly higher in IVF compared to that in IUI-OS (4 RCTs, 908 couples, 50.3% vs 43.2%, HR 1.10, 95% CI 0.54 to 2.21, I2  =  68.7%). For the safety primary outcome, the rate of multiple pregnancy was not significantly lower in IVF than in IUI-OS (3 RCTs, 923 couples, 3.8% vs 5.2%, OR 0.78, 95% CI 0.41 to 1.50, I2  =  0.0%). Clinical pregnancy (4 RCTs, 933 couples, OR 1.09, 95% CI of 0.78 to 1.53, I2  =  14.3%) and pregnancy loss (3 RCTs, 760 couples, OR 0.97, 95% CI 0.55 to 1.72, I2  =  0.0%) were comparable between IVF and IUI-OS. There were no significant differences in neonatal outcomes between the two interventions on gestational age and birth weight. Limitations, reasons for caution Four RCTs did not share IPD which may introduce the risk of data availability bias. Only two included RCTs collected data on neonatal outcomes. Three of the included RCTs predominantly or only included couples with poor prognosis of natural conception which limits generalisability. Wider implications of the findings IVF and IUI-OS are both viable options in terms of effectiveness and safety for managing unexplained infertility, especially for those with a poor prognosis of natural conception. The associated costs of interventions and the preference of couples are important in clinical decision-making. Trial registration number not applicable

Study question Can in couples with unexplained infertility a prognosis-tailored management strategy, that delays treatment if natural conception prospects are good, reduce costs without affecting live-birth rate? Summary answer In couples with unexplained infertility, use of a prognostic tool for natural conception followed by expectant management in good-prognosis couples is cost-effective. What is known already Few countries have guidelines for the assessment of the likelihood of natural conception to determine access to publicly funded ART. In the Netherlands and New-Zealand, couples with unexplained infertility who have a good prognosis for natural conception are encouraged to delay starting ART. However, the cost-effectiveness of this prognosis-tailored treatment strategy has not been determined. Study design, size, duration We studied couples with unexplained infertility to compare a prognosis-tailored strategy to care-as-usual. In the prognosis-tailored strategy, couples were assessed using Hunault’s prediction model. In good-prognosis couples (12-months natural conception >40%), outcomes without ART were modelled by censoring observations after start of ART. We then assumed that poor-prognosis couples (<40% natural conception) were treated, while good-prognosis couples delayed the start of treatment for 12 months. Data for the care-as-usual model were based on real observations. Participants/materials, setting, methods We studied 272 couples with unexplained infertility. Costs of in vitro fertilisation (IVF) and intra-uterine insemination (IUI) were calculated based on the out-of-pocket costs and Australian Medicare costs. In a cost-effectiveness model, we compared costs and effects of both strategies. Main results and the role of chance The prognostic model classified 272 couples with unexplained infertility as favourable (N = 107 (39.3%) or unfavourable prognosis (N = 165 (60.7%)) for natural conception. In the prognosis-tailored strategy, the cumulative live-birth rate was 71.1% (95% CI 64.7% - 76.4%) while the number of ART cycles was 393 (353 IVF; 40 IUI). In care-as-usual strategy, the cumulative conception rate leading to live-birth for the cohort of 272 couples, who underwent a total of 398 IVF cycles and 48 IUI cycles, was 72.1% (95% CI 65.7% - 77.4%). Mean time to conception leading to live birth was 388 days in the prognosis-tailored strategy and 419 days in the care-as-usual strategy. This translated for the 272 couples into potential savings of 45 IVF cycles and eight IUI cycles, which cost a total of AUD$ 125,817 for out-of-pocket and AUD$ 264,497 for Australian Medicare. The average cost savings per couple was AUD$ 1,435 (out-of-pocket AUD$ 463 per couple and Australian Medicare AUD$ 962 per couple). The incremental cost-effectiveness ratio, which was calculated as the total costs per additional live-births, was AUD$ 143,497 per additional live birth. Limitations, reasons for caution This study was limited to couples at a single IVF clinic. The modelling was also based on several key assumptions, particularly the number of fresh and frozen embryo transfer cycles for each couple. Wider implications of the findings: Our results show that in couples with unexplained infertility the use of a prognostic model guiding the start of an IVF-treatment reduces costs without compromising live birth rates. Trial registration number Not applicable

Study question In women with threatened miscarriage, does progesterone supplementation increase the probability of live birth? Summary answer In women with threatened miscarriage, 400 mg progesterone nightly, from onset of bleeding until 12 weeks, did not increase live birth rates. What is known already Women with a history of miscarriage who present with bleeding in early pregnancy may benefit from the use of vaginal micronized progesterone 400 mg. A recently published large randomised clinical trial indicated no overall benefit for progesterone until 16 weeks, although subgroup analysis in women with bleeding and at least one previous miscarriage, progesterone might be of benefit (Coomarasamy et al; N Engl J Med 2019;380:1815-1824). Study design, size, duration We performed a single centre placebo-­controlled randomised clinical trial. After informed consent, women with threatened miscarriage as apparent from vaginal bleeding under 10 weeks, were randomised to 400 mg vaginal micronized progesterone or placebo. The primary endpoint was livebirth. Secondary endpoints were perinatal outcomes, including preterm birth and birthweight. The planned sample size was 386 women. At a planned interim analysis randomisation was halted at 278 women due to lack of effectiveness and slow recruitment. Participants/materials, setting, methods Between February 2012 and April 2019 we randomised 139 women to 400 mg vaginal micronized progesterone and 139 women to placebo. Primary outcome data are available for 134 women in the progesterone arm and 130 women in the placebo arm. Mean age was 30.7 and 30.4 years. The number of women without a previous miscarriage was 68 (51%) and 55 (42%), while 66 (49%) and 75 (58%) women had at least one previous miscarriage. Main results and the role of chance The live birth rates were 113/134 (84.3%) and 112/130 (86.2%), respectively (RR 0.98, 95% CI 0.89-1.08). Among women with at least 1 miscarriage live birth rates were 55/66 (83.3%) and 65/75 (86.7%) (RR 0.96, 95% CI 0.84-1.11). The number of women with more than 1 miscarriage was limited (26 vs 33 in total), but no effect was seen from progesterone in these women. Preterm birth rates were 12.9% and 9.3% (RR 1.38; 95% CI 0.69 to 2.78). There were five pregnancy losses between 20 and 23 weeks, all in the progesterone arm. Mean birth weight was 3310 vs 3300 gram (p=.99). There were also no other differences in obstetric and perinatal outcomes. Anxiety, stress and depression scores did not differ between the groups. Limitations, reasons for caution Our study was single centre and did not reach the planned sample size. We stopped study medication at 12 weeks which might explain the difference between our study and studies that continued progesterone till 16 weeks. Wider implications of the findings In women with threatened miscarriage, 400 mg vaginal progesterone did not improve live birth rates. Trial registration number ACTRN12611000405910