1Introduction

Background

Uterine leiomyomata, or fibroids, are benign tumors of the uterus involving smooth muscle, or myometrium, and extracellular matrix proteins -- collagen and elastin. The annual incidence of diagnosed fibroids in one prospective U.S. cohort of women aged 25 to 44 was 12.8 per 1,000 woman-years (Marshall, Spiegelman, Barbieri, et al., 1997). Symptoms attributable to fibroids are the most common indication for hysterectomy in the United States, accounting for 33.5 percent of all hysterectomies, or more than 140,000 procedures annually (Lepine, Hillis, Marchbanks, et al., 1997; Nationwide Inpatient Sample [NIS], 1997). At least 37,000 myomectomies, surgical procedures in which only the fibroids are removed, are performed annually (NIS data).

Despite the frequency with which fibroids are diagnosed and treated, there remains considerable uncertainty and controversy among clinicians and patients regarding the best way to manage fibroids, as reflected in regional variation in hysterectomy rates (Lepine, Hillis, Marchbanks, et al., 1997), variation in compliance with expert or professional organization recommendations (Broder, Kanouse, Mittman, et al., 2000), and a proliferation of sites on the World Wide Web offering information -- often with little documentation of evidence -- on hysterectomy outcomes and alternative treatments for fibroids.

Scope and Purpose

Prevalence and Incidence of Uterine Fibroids

Estimating the overall prevalence of fibroids in the population is difficult, since estimates will vary depending on the population examined, whether asymptomatic women are included, and the sensitivity and specificity of the methods used to detect fibroids. In pathological studies of hysterectomy specimens performed for all indications, fibroids have been detected in 45 percent (Ojeda, 1979) to 77 percent (Cramer and Patel, 1990) of specimens. In the study by Cramer and Patel, the prevalence of fibroids in routine pathological examination was 76 percent in women with a preoperative diagnosis of fibroids and 46 percent in women without such a preoperative diagnosis. When serial sections were obtained at 2 mm intervals, small areas of fibroid development were detected in 77 percent in both groups, although the number and size of fibroids was greater in women with a clinical diagnosis of fibroids (Cramer and Patel, 1990).

The prevalence of clinically detectable fibroids in asymptomatic women appears to be much lower. A recent Scandinavian study using ultrasound in a random sample of 335 asymptomatic women aged 25-40 found an overall prevalence of 5.4 percent, with the prevalence increasing with age (3.3 percent in women aged 25-32 vs. 7.8 percent in women aged 33-40) (Borgfeldt and Andolf, 2000). In a series of 6706 women undergoing ultrasound examination during pregnancy, the prevalence of fibroids detected by ultrasound was 1.4 percent (Rice, Kay, and Mahony, 1989); a similar series of more than 12,000 Italian women reported a prevalence of 3.9 percent (Exacoustos and Rosati, 1993). Differences in racial distributions may play some role in the low estimates in the Italian and Scandinavian populations, although the population in the study of Rice and colleagues had large numbers of black women. We could not identify any prevalence data in nonwhite, nonpregnant populations. Lower prevalences in healthy pregnant populations may also reflect selection bias, since fibroids are associated with infertility and pregnancy complications.

Incidence is even more difficult to estimate. Most available sources of data are hospital-based, so determining the number of new cases of symptomatic fibroids is limited by lack of data on outpatient visits. Data from the Nurse's Health Study cohort showed a crude incidence rate of 12.8 per 1000 woman-years in women aged 25-44; incidence varied significantly by age and race (Table 1). Converting these data into transition probabilities in a Markov model (see Chapter 4), we estimated the cumulative risk of a diagnosis of fibroids between ages 25 and 44 to be approximately 30 percent and the cumulative risk of a hysterectomy for fibroids to be 7 percent.

Table

Table 1. Effects of age and race on incidence of fibroids.

Fibroids appear to be significantly more common in black women than in white women. Using data from the Nurse's Health Survey, Marshall and colleagues determined that the incidence rate for uterine fibroids among black women is approximately three times that of whites (Marshall, Spiegelman, Barbieri, et al., 1997). In addition, they found that black women have uterine fibroids diagnosed earlier than their white counterparts, with the highest incidence rate of diagnosis being between age 35 and 40, versus 40-44 years for whites. A study by Kjerulff et al., corroborated this finding, reporting a mean age of diagnosis for blacks of 37.5 ± 7.9 versus 41.6 ± 6.6 for whites (Kjerulff, Guzinski, Langenberg, et al., 1993).

Biology of Fibroids

Fibroids appear to arise from a single progenitor cell and are monoclonal; each individual fibroid represents a unique population of cells (Speroff, Glass, and Kase, 1999). These genetic changes affect tissue responses to both estrogens and progestins (Andersen and Barbieri, 1995; Rein, Barbieri, and Friedman, 1995). The inciting event or events for these changes are unknown. Molecular changes appear to be related to biological behavior; for example, increasing cytogenetic abnormality is correlated with increasing fibroid size (Rein, Powell, Walters, et al., 1998).

Epidemiological evidence also supports an association between estrogen, progesterone, and fibroid growth. Reproductive characteristics associated with the development of fibroids in the Nurse's Health Study include early menarche, low parity, late age at first birth, infertility, years since last birth, and early use of oral contraceptives (ages 13-16 years compared with later ages) (Marshall, Spiegelman, Goldman, et al., 1998). Increasing risk was also associated with increasing adult body mass index (Marshall, Spiegelman, Manson, et al., 1998). Some of these factors are associated with increased exposure to endogenous estrogen. However, some of these associations (e.g., late age at first birth, infertility, or low parity) may be the consequence of fibroids rather than markers for excess estrogen. Others may be the result of bias -- for example, young women may experience bleeding because of early development of fibroids, which leads to the prescription of oral contraceptives, which may then lead to an apparent causative association between early use of oral contraceptives and fibroid development.

There is also increasing evidence that bleeding related to fibroids may be the consequence of dysregulation of growth factors involved in angiogenesis (Stewart and Nowak, 1996), suggesting that medical therapies targeted at the underlying molecular mechanisms involved in bleeding may prove more effective with fewer side effects than conventional treatments.

Strategies for the medical treatment of fibroids have focused either on control of symptoms or on the manipulation of estrogen or progesterone. Use of these treatments for women with fibroids is often based on extrapolation from data on women without fibroids. NSAIDs may improve symptoms related to bleeding or pain by interfering with prostaglandin synthesis. These agents are effective in treating dysmenorrhea and menorrhagia in women without fibroids (Lethaby, Augood, and Duckitt, 2000). Oral contraceptives also reduce menstrual flow and decrease dysmenorrhea in women without fibroids by preventing ovulation and limiting endometrial proliferation, although evidence for their effectiveness in treating menorrhagia in women without fibroids is limited (Iyer, Farquhar, and Jepson, 2000). Progestins given either during the luteal phase or throughout the cycle reduce menstrual blood loss compared with placebo in women without fibroids (Lethaby, Irvine, and Cameron, 2000). Treatment with long-acting progestins, such as depot medroxyprogesterone acetate, often results in amenorrhea in women without fibroids. GnRH agonists result in amenorrhea and decreased levels of estrogen and progesterone in most patients and have been shown to result in amenorrhea and decreased fibroid size in women with fibroids (Lethaby, Vollenhoven, and Sowter, 1999).

Among invasive therapies, uterine artery embolization reduces the blood supply to fibroids, resulting in decreases in size. Myomectomy removes the fibroid itself while preserving the uterus. The fibroids may be accessed either through the cervix via a hysteroscope, or abdominally with a laparoscope or by laparotomy. While many factors influence the surgical approach (including physician and patient preferences), size, location, and number of fibroids are primary considerations. Fibroids are traditionally classified based on their location in relation to the uterine wall as either pedunculated, subserosal, intramural, or submucosal. Pedunculated fibroids are attached to the uterus by a connective tissue stalk; they may be either intra-abdominal or in the uterine cavity. Subserosal fibroids lie just underneath the parietal peritoneum (or serosa) covering the uterus. Intramural fibroids are those where the bulk of the fibroid lies within the myometrium. Submucosal fibroids have all or a substantial part of their surface immediately beneath the endometrium. The hysteroscopic approach is usually reserved for submucosal or intracavitary pedunculated fibroids. An abdominal procedure using either laparotomy or laparoscopy to gain access to the peritoneal cavity is usually used for subserosal or intramural fibroids. Hysterectomy removes the uterus along with the fibroids. Hysterectomy can be performed vaginally, through a laparotomy (abdominal hysterectomy), or through laparoscopic incisions, employing endoscopic techniques. Again, multiple factors play a role in the choice of approach, but uterine size, fibroid size, and fibroid location are the primary factors for most surgeons.

Burden of Illness

Fibroids may be asymptomatic or associated with a variety of symptoms, which are usually related to the location, size, and number of fibroids:

• Heavy or prolonged menstrual bleeding: In the Maine Women's Health Study, more than 40 percent of the women undergoing hysterectomy for a primary indication of fibroids reported bleeding for more than 8 days per month and/or heavy bleeding lasting more than 4 days per month (Carlson, Miller, and Fowler, 1994a).
• Pain: Pain associated with fibroids can be either cyclic (experienced as dysmenorrhea) or noncyclic (including primarily symptoms associated with pressure from an enlarging uterus). Forty-five percent of women undergoing hysterectomy in the Maine Women's Health Study reported pelvic pain for more than 8 days per month, and 74 percent reported discomfort (Carlson, Miller, and Fowler, 1994a).
• Infertility: Fibroids, especially submucosal fibroids, may be associated with infertility or early pregnancy loss.
• Pregnancy complications: Pregnant women with fibroids may be at higher risk for pre-term labor, placental abruption, and other pregnancy complications; risk for cesarean section is also increased (Coronado, Marshall, and Schwartz, 2000; Rice, Kay, and Mahony, 1989).

Symptoms associated with fibroids can have a significant impact on quality of life. Scores on standardized instruments for measuring quality of life are frequently significantly decreased among women with symptomatic fibroids, especially those undergoing hysterectomy (Carlson, Miller, and Fowler, 1994a, 1994b; Kjerulff, Langenberg, Rhodes, et al., 2000; Rowe, Kanouse, Mittman, et al., 1999).

Data on nonmedical costs associated with fibroids, and on costs associated with outpatient management, are scant (see Chapter 3, discussion of Question 7). However, estimated annual mean charges per admission for inpatient care in the United States for 1992 were $7,150 (Zhao, Wong, and Arguelles, 1999). In 1997, there were an estimated 231,718 admissions nationally, with mean inpatient charges of $9,041, and median charges of $8,106 (NIS data), with a total of $2.1 billion. Data comparing these numbers to other procedures commonly performed in women from the 1997 NIS data are shown in Table 2.

Table

Table 2. Number of admissions and hospital charges for procedures commonly performed in women.

Patient Populations

The primary population considered in the evidence report is women with uterine fibroids presenting with symptoms such as abnormal bleeding; dysmenorrhea or cyclic pelvic pain; noncyclic symptoms such as pressure, urinary or bowel symptoms, and low back pain; infertility; and complications of pregnancy. In addition, the report examines the available evidence on the management of women with asymptomatic fibroids.

Practice Settings

The principal practice settings considered were offices of obstetrician/gynecologists (ob/gyns), offices of other primary care providers, ambulatory surgical centers, interventional radiology suites, and acute care hospitals (for inpatient surgical procedures).

Target Audiences

The primary target audience for the report is practicing ob/gyns, who represent the majority of physicians providing care for women with symptomatic fibroids. Secondary audiences include other primary care providers; interventional radiologists; policymakers at the government, payer, integrated delivery system, and hospital levels; and patients with uterine fibroids.