Women's Imaging
Commentary
History of Uterine Artery Occlusion and Subsequent Pregnancy
Audio Available | 
Share
OBJECTIVE. During mature placenta formation, select fetal trophoblasts invade maternal decidual arterioles and junctional zone arteries and change them into low-resistance uteroplacental arteries. Consequently, physicians performing vascular procedures that occlude the uterine arteries should understand their effects on miscarriage rates, the various toxemias of pregnancy, gross and microscopic placental anatomy, and invasive placentation. Thus, the objective of this article is to review the effects of vascular occlusion on pregnancy.
CONCLUSION. In the long run, placental abnormalities may be the canary in the mine shaft to globally judge the effect of uterine artery embolization on pregnancy.
Keywords: fibroids, embolization, leiomyoma, pregnancy, uterine artery embolization
| Introduction |   |
|---|
Many authors have investigated the effect of interrupting the blood supply to the uterus on subsequent pregnancy. For a woman to become pregnant and experience a normal pregnancy after a pelvic vascular occlusion, at a minimum, seven reproductive functions must remain intact: ovulation; menstruation; implantation and early (venous) placenta formation; mature (arterial) placenta formation; low arterial resistance in the uteroplacental unit; fetal nourishment; and myometrial control of labor, delivery, and puerperal blood loss.
At least one ovary must continue to produce oocytes; enough endometrial lining must exist to accept a blastocyst; and once implanted, the blastocyst must develop a vascular connection between mother and child. In the first trimester, the placenta is fed primarily by endometrial gland secretions and venous blood. To develop a mature placenta, one supplied by arteries, in the second trimester select trophoblast cells of the placenta must invade uterine tissues to about the depth of the junctional zone and no further. If invasion is shallow, an insufficient placenta will develop. If invasion progresses deeper into the myometrium of the uterus, an invasive placenta (accreta, increta, or percreta) will develop, the placenta will not detach normally after childbirth, and postpartum hemorrhage may ensue. During this invasion, these migratory placental cells engulf and enter maternal spiral arterioles and maternal radial arteries and transform about 200 vessels into large, low-resistance uteroplacental arteries. This transformation is referred to as “physiologic change.” If enough change does not occur in the uteroplacental arteries, resistance in the uterine arteries will not drop, as it should, and the placenta will be hypoxic [1–6].
A host of toxemias of pregnancy follow including preeclampsia and eclampsia [7–13], maternal hypertension [14], low birth weight and short gestation length [15], and low placental volume [16]. Fetal cells must be sufficiently nourished by the placenta to develop a normal-weight baby at term. After implantation, myometrium must be sufficiently intact for the junctional zone and outer two thirds of the myometrium to differentiate into a more muscular third phenotype to inhibit labor until term and then, at term, initiate labor, push the baby through the birth canal, and continue contractions for several hours after placental delivery to slow blood loss while clotting occurs within the uterine vasculature as well as to redifferentiate into a junctional zone and outer myometrium.
Menstrual history, follicle-stimulating hormone levels, ovarian ultrasound, hysteroscopy, endometrial biopsy, and endometrial ultrasound are used to assess ovulation, menstruation, and implantation. In the end, pregnancy trumps all of these clinical measurements and proves adequacy of these three functions. Early placenta formation and mature placental depth can be followed up with obstetric ultrasound or MRI. Serial obstetric vital signs and clinical laboratory tests, including soluble fms-like tyrosine kinase-1 [17–24] and uterine artery Doppler velocitometry, reflect the adequacy of physiologic change of the uteroplacental arteries. Spontaneous abortion, premature delivery, or small-for-date babies suggest inadequate nourishment.
Throughout pregnancy, fetal nourishment can be judged with serial obstetric ultrasound metrics compared with nomograms. Placental development can be assessed by detailed gross and histologic examination of the placenta after delivery. Premature or inadequate labor and delivery and postpartum hemorrhage each suggest abnormal myometrium. An optimal investigation into the effect of vascular occlusion on future pregnancy should, at a minimum, collect these outcome variables. As we shall see, most studies focus on only one or two of these, primarily the ability of a woman to become pregnant and deliver a viable baby.
Why would anyone want to occlude the blood flow to the uterus in the first place? Initially, vascular occlusion of pelvic arteries was performed to control bleeding during or after childbirth. Later, occlusion therapies were developed to palliate or treat a wide variety of pelvic disorders. This article is a review of the effects of vascular occlusion on pregnancy.
| Methods of Vascular Occlusion | |
|---|
Methods for occluding the blood supply to the uterus have been around for a long time. In a study by Kelly [25] in 1894, hysterectomy, alone, was not sufficient to stop pelvic bleeding. Ultimately, the internal iliac arteries had to be ligated. However, as late as the 1940s, anatomists and clinicians still insisted that after internal iliac ligation, the bladder and other pelvic organs would slough because of inadequate blood flow [26, 27]. In 1954, like Kelly, Hecht and Blumenthal [28] reported the successful control of vaginal bleeding by internal iliac ligation after hysterectomy. Subsequently, Sack [29] described the successful use of the same technique to treat massive postpartum hemorrhage in two women after forceps deliveries. In these two cases, pelvic bleeding was controlled without hysterectomy, a large step forward in obstetric care.
Because internal iliac ligation diminishes but does not stop blood flow in the uterine arteries, direct uterine artery occlusion evolved [27, 30–33]. In 1952, Waters [26], in a study that was widely read around the world, described selective uterine artery ligation to treat postpartum hemorrhage in eight women. He stated, “Uterine artery ligation will effect the same end as hysterectomy, if the bleeding is from so-called `atony.'” He used chromic catgut as the ligature material and knowingly did not divide the uterine arteries to allow future recanalization of the ligated arteries. With this technique, each main uterine artery was ligated before its division into cervical and ascending uterine artery branches. In 1962, Aleksandrov [34] expanded Waters's technique to include bilateral uterine artery ligation and ligation of one of the uteroovarian communicating arteries to control postpartum hemorrhage in two women. The technique was further expanded to include the bilateral ligation of the uterine and uteroovarian communicating arteries to treat postpartum hemorrhage [35]. Tsirul'nikov [36] expanded the technique even further with bilateral ligation of the uterine arteries, the uteroovarian communicating arteries, and the arteries of the round ligament to treat postpartum hemorrhage. The surgical literature showed that in postpartum women, ligation of all visible arterial input to the uterus did not damage the uterus. Many reports of successful treatment of postpartum hemorrhage with uterine artery ligation have been published [37–46].
In 1964 Bateman [47], a student of Waters, ligated and then resected a segment of each uterine artery from three women who suffered from menorrhagia. In follow-up varying from 6 to 12 months, each woman returned to normal menstrual cycles. Not only was the surgical ligation and division of the uterine arteries clinically successful, Bateman's work showed that in women who were temporally distant from childbirth, the uterine arteries could be separated from the uterus without significant uterine damage.
In the mid-1970s, pelvic arterial embolization was introduced as an alternative to surgery to palliate pelvic bleeding caused by malignancies [48–54]. Since then, many have described the successful use of pelvic artery embolization to treat a wide variety of clinical conditions associated with vaginal bleeding [55–73].
In 1979, in two separate articles regarding the same case, Heaston et al. [74] and Brown et al. [75] described angiographically directed pelvic arterial embolization using “absorbable gelatin sponges” to control postpartum hemorrhage in a 22-year-old woman whose surgical blood loss could not be controlled with the combined therapies of bilateral internal iliac artery ligation, total abdominal hysterectomy, and vaginal packing. After these landmark articles, others reported successful selective embolization of various pelvic arteries to control postpartum hemorrhage [61, 63, 64, 76–101].
Uterine artery embolization (UAE) as a treatment for women with symptomatic fibroids was discovered in 1995 [102, 103]. After initially using UAE to diminish blood loss due to myomectomy, Ravina et al. [104] later proposed UAE as a primary treatment of women with symptomatic fibroids. After this group's initial small case series, UAE was widely adopted and studied around the world [105]. Ravina et al. believed that UAE killed fibroids and spared myometrium because, “Physiologically, the blood flow goes preferentially from the uterine artery toward the peripheral myomata [fibroid arterial] network. This also explains why embolization is easy and safe if performed in a free flow, since the particles will preferentially reach the myomata vascular peripheral network, thus avoiding the unwanted targets such as the normal myometrium or ovary.” CT and MRI studies before and after UAE and pathology studies of myomectomy and hysterectomy specimens after UAE support the alternative view that particles injected into the uterine arteries reach all the tissues fed by the uterine arteries including endometrium, myometrium, and ovaries [106, 107].
| Effects of Vascular Occlusion on Pregnancy | |
|---|
Fragmentary information about the effects of each of these forms of pelvic arterial interruption on pregnancy has been published. By 1964, three successful pregnancies after bilateral occlusion of the internal iliac arteries had been reported [108, 109]. In 1969, Mengert et al. [110] added five cases to the world's literature; the patients in three of these cases had bilateral ligation of the ovarian communicator arteries as well. Over the next quarter century, additional cases were added to the world's literature [111–115]. Undoubtedly, other successful pregnancies after internal iliac ligation exist; however, no systematic large-scale study of pregnancy after internal iliac ligation has been published. We know from what has been published that women can get pregnant after bilateral ligation of the internal iliac arteries. We do not know the pregnancy or pregnancy complication rates.
The same limitations are true of our understanding of fertility after selective ligation of the uterine arteries. Between 1964 and 1980, nearly two dozen full-term, successful pregnancies in women who had uterine artery ligation were published in the world's medical literature [38, 116–120]. From this, we can conclude that after selective ligation of the uterine arteries, pregnancy is possible. However, again, we lack the pregnancy and pregnancy complication rates.
The world's embolization literature is only a bit more complete than the surgical counterpart. Stancato-Pasik et al. [121] summarized their experience between 1990 and 1995. During these 6 years, 17 women 20–24 years old had “gelatin sponge pledget” embolization of the uterine arteries for postpartum hemorrhage. Five of the 17 underwent hysterectomy in the intervening years. Of the 12 who did not undergo hysterectomy, 11 (91.6%) resumed menses within 2–5 months after embolization. Three patients desired to become pregnant and did so and delivered healthy babies. Similarly, Deux et al. [122] reported on pregnancy in 25 women who had postpartum hemorrhage treated with pelvic artery embolization between 1994 and 2000. In this retrospective review, all women resumed menstruation and one became pregnant. In similar chart-review studies between 1997 and 2002, Ornan et al. [95–97] identified 28 women who had pelvic artery embolization for postpartum hemorrhage. Two went on to hysterectomy. Nine women were able to get pregnant and deliver healthy babies. Descargues et al. [123] reviewed 31 women treated with UAE for postpartum hemorrhage. Four women went on to hysterectomy. Of the 27 women who retained their uteruses, gynecology records were available for 25. All women returned to normal menses. Of the nine patients who wanted to become pregnant, five were able to conceive. Two patients who did not want to become pregnant did so. Of the 10 pregnancies that occurred, four women experienced spontaneous abortions in the first trimester and six carried to term, producing healthy babies.
In 2007, Gaia et al. [124] reviewed pregnancy after arterial embolization for the treatment of postpartum hemorrhage in 113 consecutive patients. Of these women, 107 were available for follow-up to assess fertility, and 99 (92.5%) returned to normal menses. Of these 99 women, 29 desired and attempted conception, and 18 (62.1%) reported a total of 19 pregnancies. One miscarriage occurred at 12 weeks of gestation, whereas 18 pregnancies reached full term. Of these 18 women, three (16.7%) developed postpartum hemorrhage secondary to abnormal placentation. All full-term newborns were healthy. Another group has also reported placental abnormalities after gelatin sponge UAE for postpartum hemorrhage [125]. Successful pregnancy after internal iliac and uterine artery embolization for gynecologic bleeding has also been reported by many authors [115, 126–130].
| Fibroids and Fertility | |
|---|
Assessing the effect of vascular interventions and pregnancy in women who have fibroids is doubly complex because it has long been believed that fibroids themselves negatively impact fertility. In a comprehensive review of the literature, Buttram and Reiter [131] tabulated the results from nine clinical series that each evaluated the effect of fibroids on fertility and calculated that 27% of women with fibroids had fertility problems. In a review of 18 published series of pregnancy rates for infertile women after myomectomy, the same authors calculated that 40% of women become pregnant after myomectomy. Furthermore, in a review of seven series that examined the effect of fibroids on fetal wastage, fetal wastage decreased from 41% to 19% in women with fibroids treated with myomectomy [131]. However, because fibroids are quite common, these studies certainly reflect some degree of coincidence. In a meta-analysis of prior publications, Pritts [132] stratified fibroids by location and showed that fibroids had no effect on fertility if they did not have a substantial submucosal or cavitary component. When a submucosal or cavitary component was present, fibroids decreased fertility and myomectomy improved fertility. In a prospective study of 434 women requiring assisted conception, Hart et al. [133] showed that when intramural fibroids were ≤ 5 cm in diameter, pregnancy, implantation, and ongoing pregnancy rates were reduced by half. Even when intramural fibroids were ≤ 2.3 cm in diameter, their presence still negatively influenced a woman's odds of maintaining an ongoing pregnancy.
In addition to decreasing fecundity, the presence of fibroids can also complicate a pregnancy. Qidwai et al. [134] followed pregnancies in 15,104 women with routine second-trimester ultrasound examinations. Comparing the obstetric course in women with and without fibroids, they observed that fibroids are associated with increased risk of cesarean delivery, breech presentation, malposition, preterm delivery, placenta previa, and severe postpartum hemorrhage.
| Fertility After UAE for the Treatment of Symptomatic Fibroids | |
|---|
Because a large number of UAE procedures for fibroid treatment have been performed worldwide, considerable information is available about pregnancy after UAE. Only 3 years after the original publication by Ravina et al. [103], this group reported that seven pregnancies occurred in their treatment population [135]. Five of the seven women were over 38 years and two were below 30 years of age. One woman experienced a miscarriage, and a woman with AIDS delivered prematurely. Five additional pregnancies were later reported from the same group [136, 137]. McLucas et al. [138] reported pregnancy information after UAE in 400 women, 139 of whom wished to remain fertile after UAE. Seventeen pregnancies occurred, with 10 normal deliveries. Five women aborted spontaneously, and two were still pregnant at the time of publication. Many additional case reports and clinical series describing pregnancy after UAE have appeared [139–146]. One case report described spontaneous twins in a 29-year-old woman after UAE [147]. Recently, Pinto Pabón et al. [148] followed 57 women who wanted to become pregnant in a series of 100 women treated with UAE for symptomatic fibroids. Eleven (19.3%) pregnancies were observed in these 57 women, resulting in eight live births. Early miscarriages occurred in three (27.3%) of these 11 women. No abnormal placentation was observed.
Clearly, it is possible for a woman to become pregnant and deliver a healthy baby after UAE treatment of fibroids. The questions then become, at what frequency does pregnancy occur and how normal are these pregnancies? Pregnancy rates vary considerably. Gaia et al. [124] reported pregnancy in 18 (62.1%) of 29 women seeking pregnancy. Honda et al. [149] reported that five (50%) of 10 women seeking pregnancy conceived after UAE. On the other hand, Pinto Pabón et al. [148] reported a rate of only 19.3%.
A comprehensive report of pregnancy after UAE came from the Ontario Uterine Fibroid Embolization Trial [150–152]. In this trial, 555 women underwent UAE. The average patient age was 43 years with a range from 18 to 59 years. Twenty-two women conceived, three twice, for a total of 25 pregnancies. Four spontaneous and two elective abortions were observed. Eighteen live births were reported. Fourteen were at term and four were preterm. Nine women delivered vaginally and nine by cesarean delivery. Four newborns were small for gestational age. None of these statistics should raise eyebrows given that they come from a population of mothers with fibroids and an average age of 43 years. However, placentation abnormalities, which are rarely seen, were observed at a very high rate: three (16.7%) of the 18 pregnancies. Two cases of placenta previa, one of which was a placenta accreta and the other placenta accreta that required cesarean hysterectomy, were identified. All three women with placental abnormalities were nulliparous, which makes these findings of even greater concern. Additionally, two (3.4%) of 18 pregnancies were complicated by gestational hypertension, which may have been secondary to a placentation abnormality as well.
To better understand the effect of UAE on pregnancy, Goldberg et al. [153] performed a systematic literature search comparing UAE to myomectomy. They examined published materials and contacted authors for additional data to compare 53 pregnancies after UAE with 139 after myomectomy. From this retrospective search (which included patient demographics, fibroid characteristics, and pregnancy complications), they showed that women who had UAE were significantly older, higher in parity, and had larger fibroids than women treated with myomectomy (all three differences significant at p ≤ 0.001). Consequently, the two patient populations that the authors wanted to compare were not comparable. Despite this bias in favor of myomectomy (younger women with smaller fibroids), no statistically significant difference was observed in the rate of spontaneous abortion, postpartum hemorrhage, cesarean delivery, or small-for-gestational-age babies. At the p ≤ 0.05 level, preterm delivery and malpresentation were observed more commonly in the UAE group (16% vs 3% and 11% vs 3%, respectively). Because the UAE women were older and had larger fibroids, a difference of this magnitude would not be surprising.
In 2008, Holub et al. [154] reported a prospective, randomized comparison of the impact of UAE and myomectomy on pregnancy. One hundred twenty-two women who presented with intramural fibroids > 4 cm and who wanted to become pregnant were randomly assigned to either UAE or myomectomy. The rate of technical success, symptomatic effectiveness, postprocedural follicle-stimulating hormone levels, number of reintervention procedures for fibroid recurrence or regrowth, and complication rates were no different between the two procedures. Forty women after myomectomy and 26 after UAE had tried to conceive. More pregnancies (n = 33) and labors (n = 19) and fewer abortions (n = 6) were observed after myomectomy than after embolization (17 pregnancies, five labors, and nine abortions; differences statistically significant at the p < 0.05 level). The authors concluded that myomectomy led to, “superior reproductive outcomes in the first 2 years after treatment.”
Holub et al. [154] prospectively compared the effect on pregnancy of surgical occlusion of the uterine arteries in 38 women to UAE in 20 women who conceived after fibroid treatment. When the uterine arteries are occluded surgically, no embolic particles are present in the arteries and arterioles of the uterus. Pregnancies after UAE had a higher rate of spontaneous abortion (56%) than did pregnancies after surgical uterine artery occlusion (10.5%). The authors concluded that after UAE, pregnancies had “significantly increased risk for spontaneous abortion.”
| Conclusion | |
|---|
With respect to the seven reproductive functions introduced at the beginning of this article, how many did Firouznia et al. [155] evaluate in “Pregnancy After Uterine Artery Embolization for Symptomatic Fibroids: A Series of 15 Pregnancies”? Menstrual histories after UAE were not reported. What we do know is that the crude pregnancy rate in this relatively young population of women with symptomatic fibroids (mean age, 32.3 years) was high at 60.9%. One baby was small for gestational age. One postpartum hemorrhage was observed, although the reason for the postpartum hemorrhage was not established. Was it atony related to myometrial inadequacy, was it an invasive placenta, or was it related to the surgical delivery? The toxemias of pregnancy are not mentioned in this article. Did none occur? Two spontaneous abortions were recorded. No woman was allowed to deliver vaginally, so information about myometrial adequacy is unknown. Because the data for this study were obtained from “physician's offices and hospitals,” detailed pathologic examination of each placenta, unfortunately, was not available.
Given how the arterial placenta forms, we would expect to observe that embolic particles located in and around uterine arterioles and arteries might present a biologic dilemma to migrating fetal trophoblasts. Fetal trophoblasts have evolved to transform normal maternal decidual arterioles and junctional zone arteries into low-resistance uteroplacental arteries. Evolution has not prepared these migratory cells to bump into and, somehow, deal with plastic particles attached to the walls of uterine arteries. Consequently, in addition to pregnancy rates and birth weights, understanding the effects of interrupting the blood supply to the uterus should also include a focus on early spontaneous abortion, the various toxemias of pregnancy, gross and microscopic placental anatomy, and invasive placentation. In the long run, placental abnormalities may be the canary in the mine shaft for globally judging the effect of UAE on pregnancy.
This article is a commentary on “Pregnancy After Uterine Artery Embolization for Symptomatic Fibroids: A Series of 15 Pregnancies” by Firouznia et al. published in this issue of the AJR.
Address correspondence to F. Burbank ([email protected]).
| References | |
|---|
