The article “Submucosal Fibroids Becoming Endocavitary Fol lowing Uterine Artery Embo lization: Risk Assessment by MRI” by Verma et al. [1], which appears on the preceding pages of this issue of the AJR, raises important questions for radiologists and gynecologists.

After uterine artery embolization (UAE), is fibroid movement from a position relatively distant from the uterine cavity to a position adjacent to, substantially in, or even free within the uterine cavity a complication? Similarly, after UAE is spontaneous vaginal passage of a fibroid a complication? Finally, if a woman undergoes UAE and later needs cervical dilatation to allow passage of a fibroid that is free in the uterine cavity or a hysteroscopic re section to remove a protruding sub mucosal fib roid, is that a complication or a treatment failure?

The degree of “aliveness” of fibroids varies considerably among women and even among individual fibroids within a single woman [2]. Some fibroids are highly cellular and are well perfused whereas others are degenerated to the point that no living cells are seen and no perfusion is measured. In most cases, UAE acutely turns living fibroids into dead ones. Contrast-enhanced MRI is used before UAE to distinguish perfused living fibroids from necrotic fibroids. When necrotic fibroid tissue is exposed to the uterine cavity (which is not always sterile), there is significant risk of infectious complications.

The connection between fibroids and some gynecologic symptoms is not documented nor is it particularly well understood. Most women with fibroids are symptom-free whereas other women with similar fibroid burdens are debilitated. The two most common complaints of women with fibroids are bulk symptoms and menorrhagia. Bulk symptoms are easy to understand. When a uterus enlarges beyond the size of a 12-week pregnancy—by a single huge fibroid, by many small ones, or by some combination—symptoms of pelvic pressure or pain occur. As the fibroid uterus fills the pelvis, there is not enough room in the pelvis for the rest of the pelvic organs, most commonly the bladder and the rectum, and symptoms arise.

On the other hand, the connection between fibroids and menorrhagia is not straightforward. If the mass of the fibroid were a cause of menorrhagia, it would stand to reason that small fibroids would be associated with less menstrual blood than larger ones (i.e., menstrual blood loss and fibroid volume would be correlated). However, in general, studies that have measured these two variables show no statistically significant correlation. In a study of hysterectomy specimens from 43 women with fibroids and menorrhagia symptoms, LevGur [3] retrospectively divided the women into two groups based on the sizes of each uterus: those ≤ 280 g and those > 280 g. Average weight in the ≤ 280-g group was 167.1 g; in the > 280-g group, 694.2 g, a significant difference at p < 0.0001. The author incorrectly stated that 280 g is the weight of the pregnant uterus at 12 weeks' gestation [3]. It is not. At 12 weeks' gestation, the average uterus is approximately 380 g in size [4]. It would appear that the author got his “2” and “3” confused. Complaints of menorrhagia were present in 69.8% of the women with smaller uteri; 69.9%, in larger. From this observation, the author concluded that menorrhagia was not related to uterine weight. As one would expect, the data did show a connection between symptoms of pelvic pressure and uterine weight.

At two conferences, Arleo et al. [5, 6] presented the results of 80 women with fibroids treated with UAE. Uterine and dominant fibroid volumes were measured by MRI. No correlation was observed between degree of menorrhagia and starting uterine or dominant fibroid volume. Sulaiman et al. [7] measured menstrual blood loss with the reference standard alkaline hematin method and uterine volume with MRI in 50 women on a waiting list for UAE [8]. Forty-two women complained of menorrhagia and in 33 (78.6%) of these women, menstrual blood loss was greater than 80 mL per menses, meeting the definition of menorrhagia. Median uterine volume was 510.5 mL. Uterine volume did not correlate significantly with menstrual blood loss. In fact, the diameter of the largest fibroid correlated negatively with menstrual blood loss across all fibroid locations: r = –0.57, p < 0.01; r = –0.40, p < 0.01; and r = –0.84, p < 0.001 for submucosal, intramural, and subserosal fibroids, respectively. Astoundingly, as the dominant fibroid got larger, menstrual blood decreased. It is difficult to maintain that fibroids cause menorrhagia when women with a larger fibroid burden have less menorrhagia than women with less of a burden.

Furthermore, when women with symptomatic fibroids are treated with UAE, no correlation has been observed in the response to therapy between bulk symptoms and menorrhagia. In a presentation in 2001 and a publication in 2002, Spies et al. [9, 10] presented associative data from a study of 200 patients. Comparing bleeding improvement and fibroid reduction in size in the same women, they observed only a very weak association between volume change and menorrhagia improvement. DeSouza and Williams [11] presented raw data for the percentage of dominant fibroid reduction and residual menstrual symptoms in 11 women after UAE. The product–moment correlation coefficient for their data was r = 0.18, p > 0.59, which shows no significant correlation between the two variables.

In an analysis of 538 women treated with UAE, Pron et al. [12] state, “Improvements in menorrhagia were not related to initial uterine volume (p = 0.08) or post-UAE volume reduction (p = 0.11).” In a later publication including 429 patients, Pron et al. [13] restate the lack of correlation between improvement in menorrhagia and uterine volume reduction. In a study of 50 women treated with UAE whose fibroid sizes were followed with MRI and whose menorrhagia was measured by the reference standard alkaline hematin method, “...there was no relationship found between changes in MBL [menstrual blood loss] and uterine volume.” [8, 14] Finally, Huang et al. [15] reported the results of UAE as treatment of 233 women with symptomatic fibroids. Successful resolution of presenting symptoms, predominantly menorrhagia, was not correlated with uterine volume reduction. These observations imply that two independent fibroid disorders are being treated by UAE: fibroid mass effect and menorrhagia.

This, in turn, brings into question whether it is necessary for living fibroids to be present on MRI to offer UAE treatment to women whose symptoms are exclusively men or rhagia. This logic is the justification for offering UAE to women who have menorrhagia associated with adenomyosis alone and might be the future rationale for offering UAE to women with dysfunctional uterine bleeding.

Verma et al. [1] state that signs and symptoms of women with fibroids “...may be determined by location.” Others have made this assertion [16–18]. For example, without supporting evidence, Stewart [18] states that “Submucosal myomas [fibroids], those in or partially intruding into the endometrial cavity, are most likely to cause menorrhagia.” Are these assertions supported by clinical evidence?

For a healthy viable fibroid, location appears to have little to do with its effect on menorrhagia. Jacobson and Enzer [19] studied 103 consec utive hysterectomy specimens and measured fibroid location carefully. Gross and histologic study of the endometrium overlying the fibroid, on the wall opposite the fibroid, and at a distance from the fibroid were compared. No differences were observed among the three locations across the 103 specimens. When their data are cast into a 2 × 2 chi-square table, no significant difference is noted in the presence or absence of menorrhagia in submucosal fibroids versus fibroids at all other locations (p = 0.38). Using MRI, Khilnani et al. [20] studied 123 women with fibroids and divided them into those with a submucosal component and those without. No significant difference in the frequency of menorrhagia was observed between the two groups (p = 0.07). Similarly, Wegienka et al. [21] studied 596 women with fibroids and observed, “...nonsubmucosal leiomyomata [fibroids] were associated with essentially the same increase in heavy bleeding as submucosal leiomyomata of similar size.” In a study of 80 women with symptomatic fibroids treated with UAE, Arleo et al. [5, 6] divided fibroid location into those in a submucosal location or with a substantial submucosal component and those without involvement of the submucosa by MR examination. No association between menorr hagia and fibroid location was observed. Consequently, it seems more clinical myth than objective fact that submucosal fibroids are significantly associated with menorrhagia.

Unlike healthy fibroids, which may be present without causing clinical symptoms even when in a submucosal location, fibroids that die acutely—with or without a physician's help—and erode to a location adjacent to or in the uterine cavity, can cause problems. The uterine cavity is not always sterile, and necrotic tissue of any sort is easily infected. Infection spreading from the uterine cavity to the systemic venous circulation is particularly worrisome when the endometrial lining is denuded during a menstrual period and fluids in the cavity can pass freely into uterine veins [22]. Furthermore, except during pregnancy, the uterus is programmed to expel masses within in its cavity. When so stimulated, painful cervical dilatation and laborlike cramping follow.

It has been known for a long time and is well documented in the world's literature that fibroids can spontaneously acutely infarct and can spontaneously work their way into the endometrial cavity and be expelled through the vagina. Acute fibroid infarction occurs most commonly after childbirth and has been described in many reports [23–29]. Not only are fibroids infarcted after birth, some infarcted fibroids are spontaneously expelled as well. Many case reports and small clinical series document the spontaneous postpartum expulsion of necrotic fibroids [30–36]. To our knowledge, no systematic study of this phenomenon has been published; conseq uently, the rate of spontaneous fibroid expulsion after childbirth is unknown. Nonetheless, it is clearly a natural biologic process in humans.

UAE and a closely related technology, laparoscopic uterine artery occlusion (UAO), have simply allowed physicians to do what may occasionally occur naturally [37]. Fibroid expulsion has been described after laparoscopic UAO [38] and after UAE [39–46]. The rate of fibroid expulsion after UAE has been estimated to be 3–12% [39, 47].

In the UAE literature, passage of dead fibroids into the uterine cavity and then out of the uterus by expulsion was originally described as a “complication.” Given that this was almost certainly a surprise to the women who had UAE, and probably a surprise to the interventional radiologists doing the UAE, it is understandable that this was considered a complication. How ev er, in reality the erosion of a fibroid to a location close to or in the uterine cavity represents the chance for UAE to become a myomectomy. From a biologic perspective, it matters little if a fibroid passes out of the uterus spon taneously or if it needs a little help from cer vical dilators or an operative hysteroscope. Once a fibroid is no longer in a woman's body, a myomectomy has been accomplished.

Consequently, I believe that the article by Verma et al. [1] potentially tells a very positive story. If each woman who anticipated UAE were told that fibroids near the uterine cavity may actually fall free into the cavity or protrude further into it and either spon taneously pass or require cervical dilatation or hysteroscopic resection to assist passage, “UAE myomectomy” would be the message to the patient, not “complication.” Further, because after UAE some fibroids located deeper in the myometrium become partially submucosal and accessible to hysteroscopic resection, this possibility should be described before UAE as well.

In an ideal world, one with no competition between gynecologists and interventional radiologists doing UAE, the sequence of UAE followed by cervical dilatation or operative hysteroscopic resection would be seen as a positive one and not considered a complication. If each woman anticipated that cervical dilatation or operative hysteroscopy might follow UAE to complete a myomectomy, who would consider these additional steps a complication?

This article is a commentary on “Submucosal Fibroids Becoming Endocavitary Following Uterine Artery Embolization: Risk Assessment by MRI” by Verma et al. published in this issue of the AJR.

Address correspondence to F. Burbank ([email protected]).