Impact of definitive uterine artery occlusion on ovarian reserve markers in laparoscopic myomectomy: a randomized controlled trial with 2-year follow-up.

Uterine leiomyomas, also known as fibroids, are benign monoclonal tumours that arise from myometrial smooth muscle cells. Their lifetime estimated cumulative incidence is higher than 80% in women of African origin and 70% in Caucasian women ( Baird et al. , 2003 ; Giuliani et al. , 2020 ). Although most leiomyomas are asymptomatic, according to their size, number, and location, as defined by the International Federation of Gynaecology and Obstetrics (FIGO) classification system ( Munro et al. , 2018 ), they can lead to a variety of symptoms such as abnormal uterine bleeding, pelvic pain, sexual dysfunction, pressure complaints leading to gastrointestinal and/or urinary symptoms, recurrent pregnancy loss, fertility issues, and obstetrical complications ( Bulun, 2013 ). The management strategy depends on symptoms, age, wish for future fertility and/or to retain the uterus as well as the cartography of the leiomyomas. Alternatives include hormonal treatments (combined contraceptives, progestins, gonadotropin-releasing hormone agonist, gonadotropin-releasing hormone antagonists), interventional radiology (uterine artery embolization), ablative therapies (radiofrequency, microwaves, focused ultrasound [US]), and surgical procedures (myomectomy by hysteroscopy, myomectomy or hysterectomy by open surgery or minimally invasive approach) ( Donnez and Dolmans, 2016 ). In symptomatic women with subserous and intramural leiomyomas and wishing to retain their uterus for a future pregnancy, the current gold standard is surgical myomectomy. Bleeding control is the main perioperative challenge, especially in case of large and well-vascularized leiomyomas ( Dubuisson et al. , 2016 ; Dumitrașcu et al. , 2023 ). Preoperative medical therapy such as GnRH analogues have been shown to increase haemoglobin levels before surgery and to decrease uterine and myoma size, compared with no treatment or placebo. Blood loss and the need for blood transfusion are also reduced. GnRH analogues are however associated with oestrogen withdrawal symptoms such as hot flushes which limits their use ( Lethaby et al. , 2017 ). Several techniques have been developed to minimize bleeding during surgical myomectomy, focusing on three main strategies: 1. reduction of uterine arterial blood flow with methods such as the pericervical ‘tourniquet’ technique, preoperative embolization, or preventive surgical occlusion of the uterine arteries; 2. use of uterotonic or vasoconstrictive agents such as systemic administration of oxytocin, sulprostone, ascorbic acid, tranexamic acid, and intravaginal misoprostol as well as intra-myometrial injections of vasopressin or a combination of epinephrine and bupivacaine; and 3. use of surgical site haemostatic agents such as the application of gelatin–thombin matrix. A Cochrane systematic review published in 2014 details the efficiency of the above-mentioned techniques on bleeding control during myomectomy ( Kongnyuy and Wiysonge, 2014 ). Preventive occlusion of the uterine arteries has been shown to reduce intraoperative blood loss and myoma recurrence but could also improve the quality of uterine suturing and the clinical effectiveness of treatment ( Dubuisson et al. , 2015 ). Regarding the risk of devascularization of the uterus, preservation of collaterals and vascular anastomoses quickly allows revascularization of healthy uterine tissue while residual leiomyomas continue their involution by necrobiosis, thereby reducing the risk of recurrence ( Liu et al. , 2004 ). A study on Doppler vascularization before and after uterine artery occlusion (UAO) in 43 patients did not show any significant differences in terms of uterine artery resistance indexes nor in terms of myometrial vascularization homogeneity ( Lichtinger et al. , 2003 ). A major consideration hindering the use of this technique in reproductive-age women is its effect on ovarian reserve markers. Recently, an RCT showed that temporary bilateral UAO during laparoscopic myomectomy slightly decreased anti-Müllerian hormone (AMH) levels 2 days postoperatively but not 3, 6, and 12 months postoperatively ( Ji et al. , 2018 ). Moreover, in the same study, pregnancy rates were similar in both groups during a 24-month follow-up. To date, the long-term effect of definitive UAO on ovarian reserve markers has not been evaluated. Definitive UAO associated with laparoscopic myomectomy has been shown to improve symptoms and reduce the risk of recurrence. In case of transient occlusion, the risk of recurrence seems identical to myomectomy alone ( Dubuisson et al. , 2015 ). We therefore conducted a randomized controlled trial with a 2-year follow-up evaluating the effect of preventive definitive occlusion of uterine arteries on ovarian reserve markers via sequential measures of the serum AMH levels and antral follicular count (AFC) by transvaginal US. Secondary objectives included the effect on perioperative outcomes (intraoperative blood loss, duration of the surgical procedure, intraoperative complications, pre- and postoperative haemoglobin levels, postoperative complications) and long-term effects of preventive occlusion of uterine arteries on symptoms (dysmenorrhea and uterine bleeding), leiomyoma recurrence, and pregnancies.

A total of 58 women were included in the study, 29 in the intervention arm and 29 controls. Women’s characteristics are reported in Table 1 . Patients in both groups were similar in terms of age, body index, ethnicity, nulliparity, wish to become pregnant, hormonal therapy, leiomyoma number and size, preoperative haemoglobin levels, dysmenorrhea pain score, menstrual flow, and baseline serum AMH levels and AFC values ( P s > 0.162). Characteristics of the analysis sample with comparison between the UAO and no-UAO groups (N = 59). AFC, antral follicular count; AMH, anti-Müllerian hormone; PBAC, pictorial blood assessment Chart; UAO, uterine artery occlusion; VAS, visual analogue scale. Missing values on age and adenomyosis for one patient and on baseline AFC for three patients in the UAO group. Missing values on diameter of leiomyoma, menstrual flow, baseline AM, and baseline AFC for one patient and on adenomyosis for three patients in the no-UAO group. Operative characteristics are reported in Table 2 . The mean total operative time did not differ between the two groups ( P  = 0.528). Mean blood loss during surgery was significantly lower in the UAO group (mean (SD) 138 (±104) ml) than in the no-UAO group (436 (±498) ml) ( P  < 0.001). The median intraoperative blood loss was significantly lower in the UAO (median [IQR], 130 ml [50–200]) than in the no-UAO (215 ml [131.2–612.5]) group ( P  500 ml versus 32.1% in the no-UAO group ( P  < 0.001). The difference in haemoglobin levels between the pre- and postoperative time was significantly lower in the UAO (median [IQR], 12 g/l [8.8–19.5]) than in the no-UAO (19 g/l [14–31]) group ( P  = 0.016). The mean hospital stay did not differ between the two groups ( P  = 0.967). We did not record any complication in the UAO group. We recorded one complication in the no-UAO group, which was an intraoperative haemorrhage related to leiomyoma enucleation, requiring blood transfusion (complication grade 2 according to the Clavien–Dindo classification). No complication in either of the groups was reported in the first month after surgery. Surgical characteristics of the analysis sample with comparison between the UAO and no-UAO groups (N = 59). UAO, uterine artery occlusion. Missing value on difference in haemoglobin between pre- and postoperative period and occlusion time for one patient in the UAO group. Missing values on operative time, intraoperative blood loss, hospital stay, and perioperative complications for one patient in the no-UAO group. Exact Fisher. We found that serum AMH levels did not differ between the UAO and the no-UAO group during the follow-up period ( P  = 0.633), whereas there was a significant time effect between T1 and T24 ( P  = 0.011) showing a decrease in serum AMH level across time ( Fig. 2A ). The interaction between group and time was not significant on serum AMH level ( P s > 0.275). Furthermore, UAO group achieved non-inferiority to no-UAO group for AMH dosages at each time point ( Fig. 2B ). Differences between means and 95% CI (in parentheses) were as follows: at T1 −0.11 (−2.14 to 2.40), at T3 −0.25 (−2.36 to 2.21), at T6 0.81 (−2.69 to 3.84), at T12 −0.95 (−3.15 to 1.33), and at T24 1.18 (−1.95 to 3.82). As the lower-bound CIs were greater than the non-inferiority margin of −3.5 pmol/l, non-inferiority of UAO versus no-UAO was demonstrated. Antimüllerian hormone levels and antral follicular count during the 24-month follow-up. Dynamics of serum AMH levels ( A ) and AFC ( C ) across time according to groups. Data are expressed as the geometric mean (±95% CI). Difference between AMH ( B ) and AFC ( D ) baseline-adjusted means in both conditions (UAO—No UAO) and 95% (bootstrapped) CIs at each time measure. The dashed line represents the non-inferiority margin. AMH, antimüllerian hormone; AFC, antral follicular count; UAO, uterine artery occlusion. In terms of AFC ( Fig. 2C ), we did not find any effect of intervention ( P  = 0.384) and time ( P s > 0.151), and no interaction ( P s > 0.322). Moreover, confidence intervals of the difference between AFC means in both conditions crossed the non-inferiority margin at all time points, indicating that non-inferiority of UAO versus no-UAO was not demonstrated for AFC ( Fig. 2D ). However, it appears that these confidence intervals are very large spreading in both positive and negative directions, which suggests much variability and less reliability of this measure. Regarding other clinical variables, menstrual flow decreased between baseline (T0) and all postoperative follow-up consultations ( P s < 0.046) regardless of which group patients had been allocated to ( Fig. 3A ). Moreover, the evolution of dysmenorrhea followed the same trend in both groups showing a decrease in pain between baseline (T0) and all postoperative follow-up consultations ( P s<0.001) ( Fig. 3B ). Regarding the persistence of leiomyoma, we did not find any effect of group, but there was a significant time effect between T1 and T24 ( P  = 0.033), and no interaction ( Fig. 3C ). The maximal diameter of leiomyomas increased from T1 to T24 ( P  = 0.015) but did not differ between the two groups ( Fig. 3D ). Bleeding, pain symptoms, and myoma persistence during the 24-month follow-up. Proportion (±95% CI) of PBAC scores <100 ( A ), dysmenorrhea (median and interquartile range, B ), proportion (±95% CI) with leiomyoma persistence ( C ), and leiomyoma maximal diameter (mean ± SD) across time and groups ( D ). PBAC, pictorial blood assessment chart; UAO, uterine artery occlusion. Regarding the occurrence of pregnancy, there was no difference between the two groups ( P  = 0.49). A total of 10 natural conceptions occurred during the 2-year postoperative follow-up leading to eight live births (five in the no-UAO group and three in the UAO group), one termination of pregnancy for psycho-social reasons and one first-trimester miscarriage. All deliveries were term live births. Three newborns were delivered vaginally and five were electively delivered by c-section with only one c-section for a placenta previa. There were no cases of placenta accreta. The time interval between surgery and the last follow-up visit (T24) was slightly shorter in the UAO group 25.2 (±3.4) months compared to the no-UAO group 25.7 (±2.8) months ( P  = 0.04). However, the time difference between the two groups did not exceed 1 month.

This was a single-blinded randomized controlled study performed between July 2015 and October 2021 at the division for Gynaecology of the Geneva University Hospitals to investigate the effect of preventive definitive UAO in women undergoing laparoscopic myomectomy for symptomatic leiomyomas on ovarian reserve markers. The Cantonal Ethics Committee of Geneva approved the study protocol (CCER number 14-129) and the trial was prospectively registered on clinicaltrials.gov on 9 July 2015 prior to the first inclusion ( NCT02563392 ). Eligible patients were: (i) women over the age of 18 years and under the age of 46 years old; (ii) those with a capacity for discernment, with ability to give informed consent, and who have signed an operative consent form, (iii) patients presenting myoma-related symptoms such as abnormal uterine bleeding, pelvic pain, infertility and/or recurrent pregnancy loss, (iv) patients presenting one or more leiomyomas with at least one leiomyoma classified FIGO 3 to 6 diagnosed either by US and/or MRI, and (v) patients with an indication to perform a myomectomy via laparoscopy (unique leiomyoma ≤9 cm or number of leiomyoma <4 and sum of diameters of leiomyomas ≤13 cm) ( Dubuisson et al. , 2016 ; Jain et al. , 2023 ). Exclusion criteria were: (i) pregnant women, (ii) patients having undergone previous uterine artery embolization, or (iii) patients with serum AMH levels <3 pmol/l. Women were recruited at the presurgical consultation for benign gynaecological conditions at the Division of Gynaecology of the University Hospitals of Geneva. Eligible women were identified at the consultation and referred to the research nurse for detailed information and signature of the informed consent form. From July 2015 to October 2021, 164 women were screened, and 58 patients accepted study participation (29 in each group) and were included in this study; 106 women either refused to participate or did not meet the inclusion criteria mostly because of undetectable AMH levels. We did not reach the objective of 30 patients per group because of recruitment difficulties that lengthened the study duration needing a premature termination. After randomization, two patients were excluded (one was diagnosed with a leiomyosarcoma and one had two leiomyoma FIGO 7 instead of a leiomyoma FIGO 6 initially diagnosed by US), and one withdrew her consent to participate. Moreover, 10 patients were lost to follow-up at various times, but 5 of these patients returned to follow-up at different time points (T) ( Fig. 1 ). Participant flow diagram . Prior to surgery, both groups had an initial serum AMH dosage (T0 AMH levels) and an initial transvaginal US to assess AFC (T0 AFC), leiomyoma cartography, and adenomyosis. USs were performed by trained gynaecologists using Voluson™ E8 or S8 (General Electrics) with a transvaginal 2D/3D probe and included a 2D/3D cartography of leiomyomas and a classification of their localization according to FIGO criteria for leiomyomas ( Munro et al. , 2018 ). Ovarian follicles between 2 and 9 mm of each ovary were measured and counted manually for the antral follicular count (Sono-Automated Volume Count was not used for the AFC assessment). The following US criteria for adenomyosis were described: anterior wall thickness of >35 mm, asymmetrical thickening of the myometrium, myometrial cysts, irregular or interrupted junctional zone, and linear striations. Patient symptoms were also evaluated preoperatively. Dysmenorrhea intensity was evaluated based on a visual analogue scale (VAS) ranging from 0 (no pain) to 10 (worst possible pain) and menstrual flow was assessed using the visual pictorial blood assessment chart (PBAC), with a score >100 considered as abnormal. A blood sample was drawn on the day of surgery or on the day before to evaluate preoperative haemoglobin levels. Participants did not receive any presurgical treatment such as GnRH analogues. Just prior to the initiation of the surgery, the surgeon opened the envelope containing the patient allocation (intervention or control); women were blinded to group allocation throughout the study (single-blinded trial). Laparoscopic myomectomy in both groups was performed according to previously described techniques and a specially designed endoscopic bag system (More-Cell-Safe; A.M.I., Feldkirch, Austria) was used for contained electromechanical morcellation ( Dubuisson et al. , 2016 ). Surgeries were performed by pre-experienced senior surgeons with an expertise in UAO. In the intervention group, access to the uterine arteries was carried out either by a ‘lateral approach’ (at the pelvic triangle) or by a direct ‘posterior approach’ (at the posterior part of the broad ligament) ( Dubuisson, 2007 ). The choice between the two approaches was left to the surgeon’s appreciation, depending on the intraoperative findings (uterine size and mobility, location and size of leiomyoma and adhesions) and the surgeon’s experience. In the ‘lateral’ approach, the uterine artery was accessed through a peritoneal incision between the round ligament and the external iliac vessels, then following the umbilical artery to finally dissect the uterine artery. In the ‘posterior’ approach, the pelvic parietal peritoneum was incised under the ovarian fossa, and then a direct dissection of the uterine artery in the broad ligament was performed. In both approaches, occlusion of the artery was performed by using a single 10 mm vascular clip, after having previously identified the course of the pelvic ureter. At the end of the procedure, the clips were left in place. In the no-UAO group, myomectomy was performed without prior UAO. First, 20 ml of a lidocaine-adrenaline solution (2 ampuls of 5 ml containing 5 µg/ml of adrenaline diluted in 10 ml of saline serum totalling to 0.05 mg of adrenaline) was injected into the uterine myometrium before incision and enucleation of the leiomyoma, using a needle transfixing the abdominal wall, after verifying that the injection was not intravascular via an aspiration test ( Mansour-Ghanaei et al. , 2022 ). At our institution, intra-myometrial injection of vasoconstrictive agents is the standard procedure for laparoscopic myomectomy without UAO. Systematic reviews have shown a significant reduction in bleeding with this technique ( Kongnyuy and Wiysonge, 2014 ). Operative outcomes (intraoperative blood loss, operative time, intraoperative complications, pre- and postoperative haemoglobin levels, postoperative complications) were evaluated. Intraoperative blood loss was strictly measured with the aspirated blood loss in the suction device at the end of the surgery. Operative time was defined by the time in between the skin incision and the last cutaneous stitch. Operative complications were notified by surgeons and classified according to the Clavien-Dindo classification ( Dindo et al. , 2004 ). On postoperative day 1, women had a blood test to evaluate postoperative haemoglobin levels. Women were then followed for serum AMH levels (collected and stored in a serum ban) and US AFC at 1 month (T1), 3 months (T3), 6 months (T6), 12 months (T12), and 24 months (T24) after surgery. During each US, myoma persistence and recurrence, defined as one or more leiomyoma of >2 cm, were documented. To better characterize leiomyoma recurrence, the difference in the number of leiomyomas between the immediate postoperative period T1 and the last follow-up consultation was evaluated (T24). All serum AMH levels were measured at the same time, at the end of the study, to avoid disparities in the analysis technique. The same kit was used for all dosages (Elecsys AMH Plus, Cobas e analyzers). AMH levels are expressed in pmol/l (Conversion: AMH in ng/ml = AMH in pmol/l ÷ 7.14). At T24, dysmenorrhea pain levels using the VAS and menstrual flow using the PBAC were evaluated and compared to T0 as to determine the evolution of symptoms. Menstrual flow was considered reduced if the PBAC was greater than 100 at T0 and then lower than 100 at T24. It was considered increased if the PBAC was lower than 100 at T0 and then greater than 100 postoperatively. We also monitored the occurrence of pregnancies and live births during the study. The standard practice in our centre is to perform a transvaginal US 3 months after surgery to assess the uterine scar for dehiscence or hematoma and to individualize the time interval between myomectomy and conception according to risk factors (localization and size of myomas, suturing layers, use of energy, effraction of the uterine cavity, postoperative infections) ( Parker et al. , 2010 ; Margueritte et al. , 2021 ). Women were randomized into two groups: the UAO group (laparoscopic preventive occlusion of uterine arteries and myomectomy) and the no-UAO group (laparoscopic myomectomy with intra-myometrial vasoconstrictor injection). The randomization list was created by a computer program using randomly permuted blocks of different sizes (2–4 and 6). Sealed, opaque, numbered envelopes were prepared, and patients were included consecutively by our research nurse. A sample size of 60 patients (UAO group, n = 30; no-UAO group, n = 30) was calculated to be able to show a difference, with a type I error of 5% and a power of 90%, of at least two-third of the SD for AMH values. Data were analysed using an intention-to-treat approach. Women remained in the group to which they were initially allocated at the time of randomization. No major protocol violation was reported during the study. Therefore, per-protocol or as-treated approaches would lead to the same findings. We report continuous variables as mean (± SD) or in cases of non-normal distribution, median, and interquartile range or geometric mean (± 95% CI). We checked normality of distribution graphically and according to the Shapiro–Wilk test. We report categorical variables as frequencies and percentages. We assessed differences in clinical and operative characteristics between the two groups using Brunner–Munzel tests for continuous variables and chi-squared tests for categorical variables (or exact Fisher tests when expected cell frequencies <5). To assess the impact of UAO on serum AMH levels across time, we ran a mixed-effects model on baseline-adjusted AMH values obtained at 1, 3, 6, 12, and 24 months (after applying a log-transformation), with group (UAO vs no-UAO), time (as a categorical variable), and an interaction term as fixed effects, and patients as random intercept. We assessed the assumptions of this model visually with plotted residuals, which were normally distributed and without extreme values. Moreover, in case of support for the null hypothesis (i.e. no significant difference between UAO and no-UAO), we decided (before database lock) to investigate non-inferiority of UAO for each time point using differences in baseline-adjusted mean AMH scores (untransformed) between both groups (UAO and no-UAO) with 95% CI. We obtained these values with linear regressions as associated estimates and 95% CI for each time point. We evaluated whether these confidence intervals crossed a non-inferiority margin of −3.5 pmol/l (0.5 ng/ml). We established this margin based on previous studies ( Anh et al. , 2022 ). We used the same statistical procedure for AFC with a non-inferiority margin of −3 follicles ( Zhang et al. , 2022 ). To assess the effect of surgery and UAO on menstrual flow (proportion of PBAC scores <100) and dysmenorrhea (VAS, continuous), we ran mixed-effects models including group (UAO and no-UAO), time (as a categorical variable including T0 to T24 follow-up), and an interaction term as fixed effects, and patients as random intercept (logistic model for PBAC scores and linear model for dysmenorrhea). Moreover, we assessed the effect of UAO and time (as a categorical variable including T1 to T24 follow-up measures) on leiomyoma persistence (0 myoma vs ≥1 myoma) and on leiomyoma maximal diameter (in mm) with mixed-effects models including patients as the random intercept. Finally, we compared the occurrence of pregnancy in both groups using an exact Fisher test. Statistical significance was assessed at a two-sided 0.05 level. We performed all analyses using R Statistical Software (version 4.2.2; www.R-project.org ).

Our findings suggest that definitive occlusion of uterine arteries can be safely used in women of reproductive age undergoing laparoscopic myomectomy. This technique does not compromise ovarian reserve markers and offers the added benefit of improved perioperative bleeding control. Incorporating UAO as a preventive measure during laparoscopic myomectomy may enhance the safety of the procedure ( Miller et al. , 2022 ).

To our knowledge, this is the first randomized controlled trial assessing the short- and long-term effects of definitive preventive UAO during laparoscopic myomectomy on ovarian reserve markers. Overall, the results of the present trial show no adverse effect of UAO on serum AMH levels. Serum AMH levels were indeed comparable at all follow-up time points between the UAO and the no-UAO groups. Moreover, we were able to demonstrate non-inferiority of UAO versus no-UAO regarding AMH levels. We opted to perform a definitive UAO in the treatment group due to its possible benefits over temporary occlusion, such as a reduced risk of myoma recurrence and a lower uterine hematoma rate ( Dubuisson et al. , 2015 ; Bedaiwy, 2019 ; Sanders et al. , 2020 ). Moreover, if definitive occlusion can be demonstrated to have no adverse impact on ovarian reserve, it is reasonable to infer that similar outcomes could be expected with temporary occlusion. Our results also show a significant decrease in blood loss without increasing operative time when preventive UAO is used. To enhance the accuracy of these results on blood loss, it is important to account for the use of intra-myometrial vasoconstrictive agents in the control group, which could have reduced the difference in blood loss between groups. As vasoconstrictive agents have been shown to reduce blood loss during myomectomies and are part of the hospitals’ standard procedures, it was judged unethical not to use any bleeding reduction technique in the control group. Our results show 0% of blood loss >500 ml in the UAO group compared to 32% in the no-UAO group, clearly demonstrating the efficacy of UAO over the use of vasoconstricting agents alone in preventing the haemorrhage risk. Our findings are coherent with Ji et al. ’s (2018) study that showed no statistical difference between groups in terms of AMH at 2 days, 3 months, 6 months, and 1 year postoperatively after temporary bilateral UAO in laparoscopic myomectomy. Our study adds to Ji et al. ’s study with a serum AMH level follow-up of 2 years after surgery. Our study was not powered to show a difference in live births or to assess safety on pregnancy outcomes. During the 2-year follow-up, 10 pregnancies occurred with eight live births, five in the no-UAO group and three in the UAO group. In 2019, Sanders et al. (2020) published a review and meta-analysis, on reproductive outcomes following UAO during myomectomy (open, laparoscopic, and robot-assisted) and found no reduction in clinical pregnancy, live-birth, or clinical pregnancy rates. Our randomized trial showed a significant decrease in intraoperative blood loss in the UAO group in comparison with intra-myometrial injection of a vasoconstrictor, without increasing operative time. Preventive UAO is well-documented in the medical literature as one technique to reduce blood loss in case of laparoscopic myomectomy with a blood loss decrease estimated between 50% and 60% ( Kim and Song, 2019 ; Zakhari et al. , 2019 ; Sanders et al. , 2020 ; Peng et al. , 2021 ; Hiratsuka et al. , 2022 ). Kim and Song (2019) showed a mean blood loss reduction 59% with simultaneous two-arterial occlusion. Menstrual flow and dysmenorrhea were reduced in both groups and there were no differences between both groups in terms of leiomyoma persistence/recurrence. Even if our study did not show a difference in terms of leiomyoma persistence/recurrence, a prospective observational study by Peng et al. (2021) showed a significant reduction in leiomyoma recurrence rate of 6.2% when UAO was used for laparoscopic myomectomy compared to laparoscopic myomectomy without UAO where the recurrence was of 25.9%. The data collection for US follow-up visits only described the presence or absence of leiomyomas and did not differentiate between leiomyoma persistence after surgery or recurrence of new myomas. This lack of precision could have influenced our results on leiomyoma recurrence. The strengths of this study lie in its randomized design, the long-term follow-up of the patients, the use of standardized tools for the assessment of dysmenorrhea and uterine bleeding, and the implementation of a reproducible standardized surgical technique for UAO. We must acknowledge certain limitations to our study. (i) The sample size was calculated to detect a clinically relevant difference of at least two-third of the SD in serum AMH levels, but we cannot exclude a smaller impact on serum AMH with a different sample size. Our sample size was however sufficient to demonstrate non-inferiority of UAO versus no-occlusion for AMH levels. (ii) Recruitment extended over a period of several years; the principal factors limiting participation were the 2-year follow-up and the use of titanium clip for occlusion which led to several refusals. The SARS-COV pandemic led to a long period during which surgical access for benign elective surgery was restricted, and inclusions were therefore delayed. (iii) Our study was carried out in a single centre and surgeries were performed by pre-trained senior surgeons with expertise in UAO. This certainly influenced the results in terms of time needed to perform UAO, total duration of surgery, and overall blood loss. The results might therefore not be generalized to surgical teams with less experience. (iv) Despite a long recruitment time, we chose not to involve other recruitment centres for several reasons; first, our study was hospital-funded and our research budget did not permit the extension to a multicentric design, and second, the present study was complex with a surgical intervention and a 2-year follow-up, and the single centre recruitment allowed homogeneity in presurgical care, surgical technique, and post-surgery follow-up. (v) We stopped the study after 58 inclusions without reaching the initially intended sample size of 60 participants due to hospital organization constraints and the reallocation of the study nurse to another project. (vi) Baseline and follow-up US scans were performed on different machines by several operators. Antral follicular counts and adenomyosis assessment, which are prone to inter-assessor variability could have been impacted. The variability shown in AFC measures could be related to the inability to show non-inferiority for AFC between the intervention and the control group. We did not use automated measurement (Sono-Automated Volume Count) of AFC which was not available on all US machines at the time of the study. (vii) In addition, Morphological Uterus Sonographic Assessment (MUSA) features of adenomyosis were published after the study initiation and were therefore not used to characterize the type and severity of adenomyosis ( Van den Bosch et al. , 2015 ). (viii) We assessed and described the presence of leiomyomas during follow-up US scans. We were, however, unable to follow each myoma individually. We could therefore not determine whether the myomas were persistent with an evolution, or recurrent. (ix) We did not monitor the use of oral contraceptives or hormonal therapy during the follow-up period which could have influenced menstrual flow, dysmenorrhea, and leiomyoma recurrence. (x) Finally, we used a lidocaine-adrenalin infiltration on the site of the uterotomy for enucleation which could have led to a reduction in the blood loss in the control arm and to a minimization of the difference in blood loss between groups. Lidocaine-adrenalin infiltration has been shown to be effective in preventing blood loss ( Mansour-Ghanaei et al. , 2022 ) and its local infiltration on the site of myomectomy should not have affected ovarian reserve markers.