Introduction

Uterus transplantation (UTx) has now repeatedly shown that women with absolute uterine factor infertility (AUFI) can achieve both genetic and gestational motherhood [1–3]. Along with enabling the woman to experience pregnancy, UTx eliminates the issues around surrogacy, where legalities greatly vary between countries. In comparison to most other types of transplants, UTx can utilize both deceased donor (DD) and live donor (LD) grafts. It is also currently the only transient transplant to remain in situ for a short period of the lifetime of the recipient, thus greatly reducing the risk of the well-described long-term immunosuppressive side effects, such as nephrotoxicity.

Causes of AUFI

The prevalence of AUFI has been estimated to be around 20,000 women of reproductive age, in a population of 100 million [4]. This type of in fertility may result from either uterine absence (congenital/surgical) or uterine abnormality (anatomic/functional) that prevents implantation of an embryo or completion of a viable pregnancy.

The Mayer-Rokitansky-Kuster-Hauser (MRKH) syndrome, a condition where the uterus is absent from birth, affects around 1:4,500 girls [5]. It is estimated that the MRKH cohort represents around 3% of women with AUFI and so far the vast majority of UTx attempts have been performed in women with this AUFI cause. The MRKH syndrome is usually diagnosed during adolescence, when the young woman is medically investigated due to primary amenorrhea. Diagnostic work-up will find absence of the vagina above the hymen and with only rudimentary uterine tissue above the vaginal dimple. A large proportion of women with MRKH have additional malformations in the urinary/renal system, with unilateral renal agenesis being the most prevalent co-malformation [6]. This may be relevant in the UTx setting, since a single kidney could have been a major underlying cause of preeclampsia that developed [7] during the pregnancy of the world´s first live birth after UTx [1].

A more common cause of AUFI is uterine absence due to a previously performed hysterectomy, which is the most frequent major gynecological surgery procedure that women undergo. Hysterectomy in women of reproductive age may be required secondary to symptomatic/large leiomyoma, cervical or endometrial cancer or required in the postpartum period due to massive obstetric bleeding from uterine atony, uterine rupture or malplacentation.

There are anatomical uterine abnormalities that often impede implantation, such as in women with hypoplastic uteri and a minor portion of women with unicornuate and bicornuate uterus. Although there is no difference in implantation rate in these latter two types of uteri with unification defects, as compared to normal uteri, the rates of first-trimester miscarriage [8] and preterm labor are greatly enhanced [8].

Other causes of uterine factor infertility that may benefit from UTx are adenomyosis or radiation injury of the uterus, with secondary repeated miscarriages/implantation failures. Another group is women with severe intrauterine adhesions that remain infertile despite repeated attempts of hysteroscopic adhesiolysis [9].

There is also the cohort of women with no signs of uterine disease on imaging or at endometrial investigation, but with repeated implantation failures in spite of multiple transfer of high quality of oocytes/embryos.

Preparatory Animal Studies on UTx

Around the turn of the century, animal-based research on UTx was initiated and multiple animal models have since then been used to improve and optimize the UTx procedure in order to develop the procedure to be as safe as possible before clinical introduction. The initial UTx research included rodents and this was followed by large domestic species, such as the sheep and pig, and subsequently by non-human primate models. Aspects of surgery, tolerability to ischemia, detection of rejection, immunosuppression and fertility, were investigated utilizing these animal models [10]. The research-based approach of investigating the UTx procedure was in concordance with the established Moore criteria [11] and the IDEAL recommendations for introduction of surgical innovations [12].

Fertility aspects of UTx have been studied in multiple animal species and the models have utilized autologous, syngeneic and allogeneic transplantations. The syngeneic and autologous UTx models were used to test surgical features such as uterine fixation, graft positioning, blood supply, anastomosis techniques and pregnancy in non-rejection settings. The allogeneic UTx models additionally investigated immunosuppression, detection and treatment of graft rejection and pregnancy with influences of both transplantation and immunosuppression.

The mouse was the first genuine UTx model to demonstrate implantation and that was initially with a model where the transplanted uterus was in a heterotopic position [13]. This syngeneic transplantation model, between inbred strains, contained vascular anastomoses of the caval vein and the aorta of the graft attached end-to-side to the abdominal parts of the aorta and vena cava of the recipient. Only a modest implantation rate was seen, most likely due to intrauterine accumulation of fluid, caused by poor drainage through the intra-abdominally positioned cervical opening. This led to the development of a model [14] that exteriorized the uterine cervix as a cervical cutaneous stoma. Each animal acted as internal control, since the native uterus remained in situ. In this model, pregnancy rates per uterus were similar in the transplanted uteri as compared to the native uteri, with offspring of normal birth-weight and growth trajectory up to adulthood. Furthermore, both male and female offspring from transplanted uteri were tested and seen to have normal fertility. It was also demonstrated that the uterus is an organ greatly tolerable to ischemic conditions, as live offspring were still demonstrated after cold ischemic conditions for 24 h [15].

In the rat model, fertility was examined both after syngeneic and allogeneic UTx. The syngeneic UTx model was with orthotopic transplantation, after left-sided hemi-hysterectomy, and with anastomoses end-to-side between the right common iliacs of the graft and the recipient. Similar pregnancy rates were observed in UTx animals compared with controls, and there was no difference in pups per pregnancy [16]. Growth trajectory, up to 60 days, was in offspring from animals of the UTx group and the sham-operated control group.

The first report of fertility after allogeneic UTx, in any species, was in the rat model [17]. Immunosuppression with tacrolimus was given in order to avoid rejection and experiments were terminated by caesarean section 2/3rds throughout pregnancy. The pregnancy rate was somewhat lower than that of the control group of sham-operated with tacrolimus. This first demonstration of pregnancy after allogeneic UTx was an essential milestone in the area of UTx.

In a follow-up study, using another allogeneic combination with greater tissue-type mismatch, [18] and also tacrolimus as maintenance immunosuppression, pregnancy rate was likewise somewhat lower in the UTx group, as compared to the two sham-operated control groups. Nevertheless, birth weights were similar in all groups and growth trajectory up until postnatal week 16 was unaltered in comparison to controls [18]. For the very first time research data indicated that allogeneic UTx may be regarded as safe, at least in a rodent model, in terms of perinatal outcomes.

The sheep UTx model is the most extensively used large domestic species in UTx research. Fertility in the sheep was investigated in both a non-rejection setting with the autologous UTx model, and also as allogeneic transplantation.

The autologous UTx model was with uterine-tubal-ovarian transplantation and end-to-side vascular anastomoses of the uterine artery, utero-ovarian vein and the ovarian artery, including an aortic patch, to the external iliacs [19]. After mating, pregnancy occurred in the majority of animals and offspring of normal sizes were delivered near term.

The allogeneic sheep UTx model involved hysterectomy with short pedicles of the uterine arteries and veins, which were divided cranially to the ureteric level [20]. A similar hysterectomy surgery was in parallel performed in the recipient and the uteri were shifted between these outbred sheep. Bilateral end-to-end anastomoses of the uterine arteries and veins, and attachment of the vaginal rim of the graft to the vaginal vault of the recipient, were accomplished. Maintenance immunosuppression was by cyclosporine. Pregnancies occurred in three out of five ewes that received fresh donor embryos. The three pregnancies resulted in one miscarriage, one ectopic pregnancy and one live birth at elective caesarean section [20]. Importantly, this was the first and so far only live birth from a large animal undergoing allogeneic UTx.

The only offspring in a non-human primate species reported up until today was one live birth after autologous UTx in the macaque [21]. In that report, two cynomolgus macaques underwent auto transplantation with unilateral preservation of the oviduct and the ovary. Menstruation resumed spontaneously in one macaque and natural mating resulted in a viable intrauterine pregnancy. The pregnancy was initially uneventful [21] but bleeding with signs of partial placental abruption occurred and a caesarean section was performed with delivery of a live offspring.

Live Donor UTx – Results

The first UTx attempt ever in modern history was performed in Saudi Arabia in 2000 [22]. An unrelated peri-menopausal woman planned for elective ovarian surgery, donated her uterus to a woman that had experienced emergency peripartum hysterectomy. However, the uterus became necrotic postoperatively and was removed 99 days post UTx, with presence of bilateral uterine vessel thrombosis. It was speculated that inadequate uterine structural support may have led to prolapse with associated tension and curving of the vessels, thereby disturbing supply and outflow of blood. No proper research preparations preceded this UTx case.

In 2012, a Swedish team initiated their first human clinical UTx trial, by means of psychological [23] and medical [24] screening of patients. Importantly, the trial was started after systematic UTx research for more than a decade [10] and followed the IDEAL recommendations for introduction of surgical innovations [12]. Nine LD cases were performed. They included eight MRKH recipients and one recipient with previously diagnosed cervical cancer, treated successfully with radical hysterectomy and pelvic lymphnode dissection. Seven of the nine donors were related (five mothers, maternal aunt, sister) to the recipient, one was a family friend and one a mother-in-law. All donors had completed previous uncomplicated pregnancies to term and five donors were postmenopausal. The retrieval surgery, performed through a sub-umbilical midline laparotomy, included dissection of the uterus together with the bilateral uterine arteries and deep uterine veins, including segments of the internal iliac vessels. The surgically most demanding part of the operation was the dissection of the uterine veins, because of large anatomical variations, firm attachments to the ureters/parametrium, their deep pelvic locations and the multiple connections between the 2–3 uterine veins, before convergence into the internal iliac vein. The duration of the retrieval surgery was 10–13 h, with the long duration being mainly due to the challenging and precise dissection of deep uterine veins [24]. The surgery of the recipient spanned 4–6 h and included bilateral end-to-side anastomosis of segments of the internal iliac arteries and veins of the graft to the external iliacs (Figure 34.1), anastomosis of the vaginal vault to the vaginal rim of the graft, and uterine ligament fixations.

Figure 34.1 The figure shows different techniques for anastomosis of the uterus at transplantation. The anterior portions of the internal iliac arteries (red) are in all techniques anastomosed end-to-side to the external iliac vessels (pink). Variations exist in the venous anastomosis techniques. In the upper panel, the left uterine vein (blue) with patch/segment of the internal iliac vein is anastomosed end-to-side to the external iliac vein (light blue). On the right side, two deep uterine veins, with the smaller one having been transected before retrieval (because of passage under the ureter) and then re-anastomosed on the back-table, are used. Both are attached to a segment of the internal iliac vein and that is anastomosed end-to-side to the external iliac vein. In the middle panel, the upper uterine veins (blue), before confluence with the ovarian veins, are used on both sides. On the right side, the upper uterine vein is anastomosed end-to-side to the external iliac vein (light blue). No deep uterine vein is used on this right side. On the left side the upper uterine vein is anastomosed end-to-side to the deep uterine vein, to increase outflow, and this deep uterine vein, with a patch of the internal iliac vein, is anastomosed end-to-side to the external iliac vessel. In the lower panel, the complete utero-ovarian veins on both sides, are used as the sole outflow section. These are anastomosed end-to-side on the external iliac veins at a more cranial position than the traditional anastomosis sites

The 6-months outcome [24] was that seven out of nine uteri were still viable. Bilateral thrombotic uterine vessel occlusion and persistent intrauterine infection were the causes of the two uterine removals. Despite worry over organ survival during the initial months, the nine recipients and partners showed positive psychological wellbeing and stability during the first post-transplantation year [25]. All donors were in good medical and psychological health at one year post-donation [26].

In September 2014, the first, ever human live birth after a UTx procedure occurred [1]. The recipient, with the MRKH syndrome, and the 61-year-old donor were close friends. There was an interim period of 12 months from UTx until embryo transfer (ET), which is in accordance with the international recommendations concerning solid organ transplants [27]. Implantation occurred at her first ET and apart from a single rejection episode at gestational week 18, the first two trimesters were uncomplicated [1, 28]. In gestational week 31+5, the mother acquired preeclampsia and a caesarean section, with delivery of a healthy boy, was performed. The second UTx baby [29] was delivered in November 2014, with a uterus donated by the mother of the recipient. Thus, the mother and her child had developed within the same uterus. During the period from 2014 to 2017, a total of eight healthy children were born from the seven women who underwent the full procedure of IVF, UTx and ET in this initial human UTx clinical trial [24]. Taken together, the take-home-baby rate of the trial is 85% and the clinical pregnancy rate is 100%.

Subsequently to the Swedish UTx trial around 40 LD UTx procedures have been performed up until the end of 2018 and with published data from the first cases in China, Czech Republic, Germany, USA and India.

Live donor UTx case number eleven in the world was performed in China in November 2015 [30]. A 42-year old mother donated the uterus to her daughter with MRKH. The UTx procedure included complete robotic-assisted laparoscopic retrieval surgery with the utero-ovarian veins being the sole venous outflow (Figure 34.1). This made the surgical dissection fairly easy but it necessitated oophorectomy, which is controversial in a woman still a decade away from menopause [31]. Menstruation occurred after two months and the graft was viable one year after transplantation [30]. So far, pregnancy has not been reported in this case.

The group in Czech Republic started a clinical trial involving up to 10 cases of LD UTx and 10 of DD UTx in 2014, with the early results of the first set of LD cases recently published [32]. Five recipients with MRKH, of ages 18–25 years, received uterine grafts from their mothers in four cases and from a mother´s sister in one case. The ages of the donors were 49–58 years, with three of them being postmenopausal but all having had normal pregnancies. In a majority of cases, only ovarian veins were used as venous outflow (Figure 34.1). This necessitated oophorectomy, but since the patients’ donors were either postmenopausal or within a few years before menopause this is not debatable. One uterus had to be removed around two weeks post UTx due to vascular thrombosis and venous outflow problems, which was partly predictable since a thrombectomy had been performed on postoperative day 5. Two of the remaining four LD recipients developed stenosis over the vaginal–vaginal anastomosis line and these were surgically corrected by vaginal approach. In one case, surgery had to be repeated and after that the patient developed a fistula between the bladder and the vagina, which was surgically corrected some months later. Initial ET attempts have been performed but there is yet no reported pregnancy.

The first LD cases in Germany and in USA were performed in 2016. The German group reported three UTx attempts, with one terminated after organ retrieval prior to transplantation [33]. The reason for aborting the case was that back-table flushing revealed extreme resistance to flow in one uterine artery and inability to flush the contralateral uterine artery adequately. The investigators predicted a very high risk of graft vessel thrombosis with insufficient uterine blood flow to enable survival of the organ after transplantation or to support a pregnancy [33]. In the other two cases, UTx were performed successfully and menstruations resumed within six weeks post UTx.

In the trial in the United States, the first three recipients lost their grafts during the initial two weeks, due to vascular complications [34]. The surgeries of the subsequent two cases were uneventful and at 3–6 months after transplantation the grafts were functioning [34]. In December 2017, one of the two surgically successful UTx cases delivered a healthy baby after becoming pregnant on her first ET [3]. In this case, the time span from UTx to ET was only 6 months, which is a questionable short time considering that the period of frequent rejections, necessitating high immunosuppression, is usually at least 10 months after transplantation.

The Indian team embarked in May 2017 on a partial conventional laparoscopic approach for LD UTx [35]. Although the team was experienced in laparoscopy there had been no preparations for this endeavor in the research setting concerning the two cases. One recipient had the MRKH and one was born with a uterus but this had severe intrauterine adhesions after a previous pregnancy that resulted in neonatal death and endometritis. The donors were mothers of ages 42 and 45 years. Ovarian veins were the sole outflow (Figure 34.1) and oophorectomies were performed. Surgery was uneventful in both cases and follow up has shown graft viability during the initial months.

Live Donor UTx – Suggested Research Topics

There are issues concerning LD UTx that need to be further studied, analyzed and optimized. One topic is proper donor selection with special reference to age. In the Swedish initial trial, the two graft losses were in two out of three donors that were above 60 years of age [24]. Moreover, the oldest donor in the first set of the United States UTx series, with three early graft failures, was 55 years and she had nearly completely occluded uterine arteries [34]. The oldest donor of the German series was 61 years [33] and that case was aborted during back-table preparation due to inadequate perfusion of the uterus. Histopathology showed extensive intimal fibrosis and early sclerosis of the uterine arteries. Therefore, donor age with postmenopausal state may be related to graft failure. However, since the first UTx baby was born from a uterus retrieved from a 61-year-old donor, it is not only age that matters [1]. Hereditary and lifestyle factors, especially those with a strong association to atherosclerosis, as well as postmenopausal state of a donor, are key issues. Consequently detailed preoperative imaging of uterine arteries and uterine perfusion are required. The optimal imaging modality for this, as well as threshold levels for uterine artery diameters and uterine perfusion, and for a good graft has to be decided through meticulous research studies.

Another issue is which venous outflow paths or combination of these (Figure 34.1) that may be suitable in a LD situation where surgical risk and long-term consequences for the donor have to be minimized and with the ureter to be protected all through surgery.

The Swedish trial [24], the German trial [33] and some cases of the Czech trial [32] used the deep uterine veins, connected to patches/segments of the internal iliac vein, as main outflow parts that were anastomosed to the external iliac veins. The advantage of this technique is related to that the anastomoses can be performed by use of a vein (internal vein segment) with a wall thickness that is comparable to the external iliac vein. This will render the anastomosis surgery easier. Moreover, the Swedish trial [24] also involved some cases where one 5–6 cm long proximal part of the upper uterine vein, before its convergence with the ovarian vein, was used as extra outflow after back-table anastomosis end-to-side to the deep uterine vein or end-to-end to a segment of the internal iliac vein.

As an alternative option, the bilateral utero-ovarian veins can be used as in the trials of China [30], USA [34], Czech Republic [32] and India [35]. The utero-ovarian veins drain both the uterus and the ovaries. These veins are usually two to three per side until they converge and drain into the cava and left renal vein. In a clinical LD situation, the advantage of using the utero-ovarian veins is the much easier and therefore faster dissection of the donor, which however has to be balanced by the negative effects of possibly injuring the ovaries or to be forced to perform oophorectomy, with secondary negative effects concerning cardiovascular health [31]. Moreover, there is concern that the utero-ovarian veins are not sufficient enough to optimally drain the uterus and to keep the graft optimal for long-term organ survival and pregnancy. There must be consideration, that the utero-ovarian veins, which in the UTx situation are typically anastomosed to the external iliacs of the pelvis, as opposed to the natural inlet higher in the abdominal vessels, would be stretched with the considerable rise of the uterine fundus during the later stages of pregnancy. This may compromise blood flow.

Today, there exists one premature live birth after exclusive use of utero-ovarian veins [3] and several term and premature live births after use of uterine veins [24]. Future UTx studies with pregnancy outcomes will be needed to address the question of which uterine venous outflow should be preferred in the LD UTx procedure.