Chapter  34:  Management of Uterine Fibroids Volume 1. Evidence Report: Evidence Report/Technology Assessment Number 34

Key Research Questions

The EPC addressed nine key research questions:

1. What are the risks and benefits of hysterectomy and myomectomy in the treatment of symptomatic and asymptomatic fibroids?

2. What are the risks associated with single vs. multiple myomectomies? (i.e., Do women with a solitary clinically apparent fibroid have different outcomes after surgical management than women with multiple fibroids?)

3. Who are appropriate candidates for each procedure?

4. What is the incidence of need for additional treatment after myomectomy or other uterus-sparing interventions?

5. Does additional treatment result in significantly increased morbidity? (i.e., Is the overall risk of adverse outcomes greater with uterus-conserving therapy because of recurrence or persistence resulting in additional therapy with associated risks, compared with immediate definitive therapy such as hysterectomy?)

6. What are the risks and benefits of nonsurgical treatment?

7. What are the costs associated with effective surgical and nonsurgical treatments? (This question was expanded to include other invasive therapies such as uterine artery embolization.)

8. Do risks and benefits differ by race, ethnicity, age, interest in future childbearing, etc.?

9. What are the effects of surgical management of uterine fibroids, especially hysterectomy, on the aging process?

Interventions Assessed

Interventions considered include:

1. No intervention ("watchful waiting").

2. Medical therapies; nonsteroidal anti-inflammatory drugs (NSAIDs), oral contraceptive pills (OCPs), progestational agents, other oral agents identified in a literature search (e.g., mifepristone, tibolone, herbal preparations), and gonadotropin-releasing hormone (GnRH) agonists (both as primary therapy and as an adjunct to myomectomy or hysterectomy).

3. Invasive therapies: uterine artery embolization, coagulation using cautery or laser, myomectomy, and hysterectomy.

1. No intervention.

2. Prophylactic myomectomy.

3. Prophylactic hysterectomy.

The EPC did not attempt to evaluate systematically the evidence on the relative benefits, risks, and costs of different technical approaches to either the diagnosis and followup of fibroids (such as clinical examination, ultrasound, or magnetic resonance imaging) or surgical procedures (e.g., comparing a laparoscopic to an abdominal myomectomy). Although these questions are clearly important, each topic in itself is large and complex enough to warrant a formal systematic review.

Patient Population and Settings

The primary population of interest is women between the ages of 20 and 55 years with symptomatic or asymptomatic uterine fibroids. Separate reviews were performed for women of different racial and ethnic backgrounds, ages (especially perimenopausal women), and plans for future childbearing.

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

Outcomes Considered

Outcomes considered varied depending on the study and the question being addressed. Data recorded on the abstraction forms included anatomical/physiological outcomes (change in uterine size, fibroid size, hemoglobin, or hematocrit); symptomatic outcomes (change in symptoms of bleeding, cyclic pain, or noncyclic pain); pregnancy-related outcomes (pregnancy rates, live-birth rates, pregnancy complications); quality-of-life measures; adverse outcomes (side effects and complications of treatment, development of new symptoms); need for additional treatment after uterus-conserving therapy; and resource use (length of stay, medical costs, time lost from work or usual activities).

Literature Sources

The primary sources of literature were the following databases (with search years shown in parentheses): MEDLINE (1975-February 2000), HealthSTAR (1975-February 2000), CINAHL (1983-February 2000), CancerLit (1983-February 2000), the Cochrane Library (Issue 3 1999), and EMBASE (1980-January 2000). Searches of these databases were supplemented by secondary searches that included e-mail subscriptions for announcements of newly published journal articles and thorough searching of the reference lists of all included articles and review articles.

The initial search was performed in MEDLINE (and then duplicated in the other databases) and limited to articles in the English language and with human subjects. A previously validated search strategy was used that identifies three subsets: (a) high specificity for randomized controlled trials (RCTs), using terms like "randomized"; (b) moderate specificity, using terms like "blinding"; and (c) low specificity, using terms such as "followup studies." All searches included the MeSH terms "uterine neoplasms," "leiomyoma," "hysterectomy," "hysterectomy, vaginal," and "surgical procedures, laparoscopic," as well as text terms (truncated) for "fibroid," ("uterine and leiomyoma"), "hysterectomy," and "myomectomy."

The overwhelming majority of RCTs identified by this initial search examined the use of gonadotropin-releasing hormone (GnRH) agonists, frequently as adjunctive treatment prior to surgery. The remainder of the citations identified were either uncontrolled case series, case series with historical or nonrandomized controls, case-control studies, or in a few instances, prospective cohort studies. A subsequent search (performed initially in EMBASE and then duplicated in MEDLINE and the other databases) was targeted to surgical interventions and included other, less robust study designs. Finally, because few of the targeted articles on hysterectomy and fibroids provided data relevant to Question 9 (on the effects of hysterectomy on the aging process), the EPC performed an additional search on hysterectomy without limiting studies to those including patients with fibroids.

Screening of Articles

Empirical study designs considered included controlled trials, prospective trials with historical controls, prospective or retrospective cohort studies, case-control studies, and medium to large case series (n = 20). Studies of these types and review articles were included if they met the following criteria: (a) the study population included women with uterine fibroids; (b) data were relevant to one or more of the key questions described above; and (c) information was presented on health outcomes, health services use, and/or health care costs for the management of uterine fibroids. Exclusion criteria were as follows: (a) the article was not original research or relevant review, (b) the patient population did not include women with uterine fibroids, (c) the study design was a single case report, or (d) the study design was a small case series with fewer than 20 subjects. For studies on the effects of hysterectomy on aging (Question 9), the EPC expanded the inclusion criteria to include studies of all hysterectomies for benign disease, even if the patient population did not include women with fibroids or results were not reported separately for patients with fibroids.

The searches yielded 1,084 articles. Abstracts from these articles were reviewed against the inclusion/exclusion criteria by five physician investigators. A team of two physicians reviewed each abstract; when no abstract was available, the title, source, and MeSH words were reviewed. Articles were included if requested by one member of the review team. At the full-text screening stage, each article was independently reviewed by two physicians, and disagreements were resolved through discussion. Articles on the effects of hysterectomy on aging were identified by a separate search and were reviewed by a single reviewer with a special interest in gynecological surgery and the aging process.

Data Abstraction Process

Teams of two physicians performed the data abstraction for articles identified by the main searches. For each included article, one physician completed the data abstraction form, and the other served as over-reader. The physician responsible for the primary abstraction also entered data on relevant outcomes and results into a Microsoft Word document. The information from the data abstraction form and the corresponding outcomes and results data were then merged into the evidence table format. The data abstraction assignments were made based on the physicians' clinical interests and expertise.

Again, data on outcomes of hysterectomy that did not explicitly focus on patients with fibroids were not subject to the same level of review. A single reviewer was responsible for selecting articles and assessing their relevance. Because the majority of these articles did not explicitly focus on women with fibroids, they were not abstracted into the evidence tables.

Criteria for Evaluating Article Quality

The majority of the evidence was not of the highest quality according to published grading systems, such as that used by the U.S. Preventive Services Task Force in its evaluation of preventive services or by the American College of Obstetricians and Gynecologists (ACOG) in its Practice Bulletins. In these and most other systems, randomized trials done with rigorous methodology are judged to be of the highest quality, followed by prospective cohort studies. Very few of the identified articles were of these types. Dismissing this literature entirely would have severely limited the investigators' ability to make any inferences at all; in fact, some study designs, such as large cohort studies or retrospective reviews of administrative data, may be better suited to certain questions, such as those concerning disease incidence or resource use. On the other hand, the quality of these studies clearly varies widely. Therefore, each study was evaluated for factors affecting internal and external validity. These criteria were:

For internal validity: Randomized allocation to treatment, and appropriate methods for randomization; adequate description of patients and controls; adequate description of length of followup, loss to followup, and dropout rates; and recognition and discussion of important statistical issues.

For external validity: Information about age, racial/ethnic background, pregnancy history, and prior surgical history; adequate description of uterine or fibroid size, fibroid location, and fibroid number; adequate description of baseline symptoms; adequate description of timing of outcomes measurement; adequate description of methods used for outcomes measurement; description of the validity and reliability of outcomes measures; adequate description or reference to clinical care provided to subjects; and use of standard, validated measures.

Additional Data Sources

The EPC used two additional data sources, primarily to address questions of costs and racial differences:

1. The Nationwide Inpatient Sample (NIS) maintained by the Agency for Healthcare Research and Quality (AHRQ). This database contains administrative discharge data from over 1,000 hospitals in 22 states, representing a stratified sample of 20 percent of U.S. hospitals. The NIS provided supplemental data on frequency and resource use for hysterectomy and myomectomy by age and race (Question 7, on cost, and Question 8, on age and race). These data from the sampled hospitals were converted to national estimates using the weighting variables provided by AHRQ. To assess differences in the proportion of white and black women undergoing each procedure, logistic regression was performed using myomectomy versus hysterectomy as the dependent variable and race as the main independent variable and controlling for age, payer type, and median income.

2. Data from Duke University Medical Center. Data on hospital costs and clinical characteristics affecting cost and complications of patients undergoing myomectomy were obtained from Duke University Medical Center. Records for patients undergoing abdominal myomectomy at Duke University Medical Center between July 1, 1993, and June 30, 1998, were identified using the hospital's cost-accounting system. Data on each patient's age, race, length of stay, insurance status, ZIP code, attending physician, and hospital costs were available from the abstracted discharge data. After the appropriate records were identified, trained abstractors reviewed the charts to collect additional clinical information. Multivariate analysis was used to determine the effects of multiple patient characteristics on outcomes using linear regression for continuous outcomes (length of stay and costs) and logistic regression for dichotomous outcomes (transfusions and complications).

Decision Model

The EPC constructed a Markov model incorporating 11 possible health states: (1) asymptomatic fibroids, defined as having clinically detectable fibroids but without symptoms attributable to their presence; (2) symptomatic fibroids, without side effects/complications of therapy; (3) symptomatic fibroids, with side effects/complications of therapy; (4) improved symptoms, with no side effects/complications; (5) improved symptoms, with side effects/complications; (6) no symptoms, without side effects/complications, defined as having symptoms prior to treatment but experiencing complete relief after treatment with no side effects or complications attributable to therapy; (7) no symptoms, with side effects/complications, defined as having symptoms prior to treatment with complete relief after treatment but with side effects or complications attributable to therapy; (8) uncomplicated pregnancy (no complications attributable to fibroids); (9) complicated pregnancy (complications attributable to fibroids); (10) menopause; and (11) death. Data needed to estimate transition probabilities were identified. Because data on many of the probabilities were unavailable, a simplified analysis comparing outcomes of hysterectomy, myomectomy, and no treatment in women of different ages was performed.

Overview

The primary purpose of the evidence report is to evaluate the evidence concerning the benefits, risks, and costs of various treatments for uterine fibroids. Three primary sources of evidence have been used:

1. Published studies of medical and surgical treatments for fibroids.

2. Hospital claims data from the Nationwide Inpatient Sample (part of the Healthcare Cost and Utilization Project [HCUP] supported by the Agency for Healthcare Research and Quality [AHRQ]).

3. Hospital chart data and cost data from Duke University Medical Center.

The report has two secondary goals: to formulate recommendations for future research on the management of fibroids and to develop the framework for a decision model that could ultimately be used as a tool for synthesizing data about the management of fibroids.

Key Research Questions

The key research questions addressed in the report were developed through consultation with AHRQ and our report partner, the American College of Obstetricians and Gynecologists (ACOG). The questions were as follows:

1. What are the risks and benefits of hysterectomy and myomectomy in the treatment of symptomatic and asymptomatic fibroids?

2. What are the risks associated with single versus multiple myomectomies? (i.e., do women with a single, clinically apparent fibroid have different outcomes after surgical management than women with multiple fibroids?)

3. Who are appropriate candidates for each procedure?

4. What is the incidence of need for additional treatment after myomectomy or other uterus-sparing interventions?

5. Does additional treatment result in significantly increased morbidity? (i.e., is the overall risk of adverse outcomes greater with uterus-conserving therapy because of recurrence or persistence resulting in additional therapy with associated risks, compared with immediate definitive therapy such as hysterectomy?)

6. What are the risks and benefits of nonsurgical treatment?

7. What are the costs associated with effective surgical and nonsurgical treatments? (This question was expanded to include other invasive procedures, such as uterine artery embolization.)

8. Do risks and benefits differ for women according to race, ethnicity, age, interest in future childbearing, and so forth?

9. What are the effects of surgical management of uterine fibroids, especially hysterectomy, on the aging process?

Our approach in addressing each of these questions was to identify and evaluate the relevant literature and supplemental data (if any); report the results; and, where evidence was lacking or methodological limitations in the available sources precluded drawing firm conclusions, identify the type of evidence needed to answer the question.

Interventions Assessed

We reviewed studies of the following treatments for symptomatic fibroids:

• No intervention ("watchful waiting")

• Medical therapies
  -- Nonsteroidal anti-inflammatory drugs (NSAIDs)
  -- Oral contraceptive pills (OCPs)
  -- Progestational agents
  -- Other oral agents
  -- Gonadotropin-releasing hormone (GnRH) agonists (both as primary therapy and as an adjunct therapy to myomectomy or hysterectomy)

• Invasive therapies
  -- Uterine artery embolization
  -- Coagulation using cautery or laser
  -- Myomectomy
  -- Hysterectomy

In addition, we reviewed the available evidence on the following strategies for managing asymptomatic fibroids:

• No intervention

• Prophylactic myomectomy

• Prophylactic hysterectomy

We specifically did not address questions concerning the benefits, risks, and costs of various methods used to diagnose fibroids (such as clinical examination, ultrasound, or magnetic resonance imaging [MRI]); we also did not specifically address questions about the risks, benefits, and costs of various technical approaches to the same procedure (laparoscopic vs. abdominal myomectomy, or abdominal vs. vaginal hysterectomy). While both questions are clearly important, the Duke team, AHRQ, and the advisory panel agreed that the amount of time and effort required to systematically review these questions precluded their inclusion in this evidence report.

Because the primary focus of our investigation was clinical management of fibroids, we did not attempt a truly systematic review of the "basic science" literature concerning the genetic, biochemical, and molecular mechanisms involved in the development and growth of fibroids, or the epidemiology of fibroids. Again, given the importance of the topic, such reviews are clearly warranted. Although we used primary epidemiological data as much as possible, we did not attempt to review the primary literature on underlying biological mechanisms.

Literature Sources

The primary sources of literature were six of the most widely used computerized bibliographical databases (search years shown in parentheses): MEDLINE (1975-February 2000), HealthSTAR (1975-February 2000), CINAHL (1983-February 2000), CancerLit (1983-February 2000), the Cochrane Library (Issue 3 1999), and EMBASE (1980-January 2000). Searches of these databases were supplemented by secondary searches that included e-mail subscriptions to the tables of contents of current medical journals and thorough searching of the reference lists of all included articles and review articles.

Search Strategy

We developed the basic search strategy using the National Library of Medicine medical subject headings (MeSH) key word nomenclature developed for MEDLINE. The same strategy was used to search HealthSTAR, CINAHL, and CancerLit. A Duke University Medical Center librarian checked the strategy and assisted with its translation to the key word structure used by EMBASE.

The initial search was performed in MEDLINE (and then duplicated in the other databases) and was limited to articles in English and with human subjects. It employed a previously validated (Dickersin, Scherer, and Lefebvre, 1994) search strategy that identifies three subsets: (a) high specificity for randomized controlled trials (RCTs), using terms such as "randomized"; (b) moderate specificity, using terms such as "blinding"; and (c) low specificity, using terms such as "followup studies." The overwhelming majority of RCTs identified by this initial search examined the use of gonadotropin-releasing hormone (GnRH) agonists, frequently as adjunctive treatment prior to surgery. The remainder of the citations identified were either uncontrolled case series, case series with historical or nonrandomized controls, or, in a few instances, prospective cohort studies. A subsequent search (performed initially in EMBASE and then duplicated in MEDLINE and the other databases) was targeted to surgical interventions and included other, less robust, study designs. Finally, because few of the targeted articles on hysterectomy and fibroids provided data relevant to Question 9 (on the effects of hysterectomy on the aging process), we performed a supplemental search on hysterectomy without limiting studies to those including patients with fibroids.

Table 4. Basic fibroids search strategy

Table 4. Basic fibroids search strategy

1	leiomyoma/
2	fibroid$.tw.
3	(uter$ and leiomyoma$).tw.
4	exp hysterectomy/
5	hysterectomy, vaginal/
6	surgical procedures, laparoscopic/
7	hysterectom$.tw.
8	myomectom$.tw.
9	or/4-8
10	co.fs.
11	dt.fs.
12	th.fs.
13	su.fs.
14	or/10-13
15	or/1-3
16	14 and 15
17	9 and 15
18	16 or 17
19	limit 18 to human
20	limit 19 to female
21	limit 20 to English language
22	limit 21 to yr = 1975-2000
23	13 and 15
24	17 or 23
25	limit 24 to human
26	limit 25 to female
27	limit 26 to English language
28	limit 27 to yr = 1975-2000
29	15 and 11
30	limit 29 to human
31	limit 30 to female
32	limit 31 to English language
33	limit 32 to yr = 1975-2000
34	28 or 33
35	15 and 12
36	limit 35 to human
37	limit 36 to female
38	limit 37 to English language
39	limit 38 to yr = 1975-2000
40	15 and 10
41	limit 40 to human
42	limit 41 to female
43	limit 42 to English language
44	limit 43 to yr = 1975-2000
45	44 or 39 or 33 or 28
46	45 or 22
47	randomized controlled trials/
48	random allocation/
49	double-blind method/4
50	single-blind method/
51	randomized controlled trial.pt.
52	47 or 48 or 49 or 50 or 51
53	animal/
54	human/
55	53 and 54
56	53 not 55
57	52 not 56
58	clinical trial.pt.
59	exp clinical trials/
60	(clin$ adj trial$).tw.
61	((singl$ or doubl$ or trebl$ or tripl$) adj (blind$ or mask$)).tw.
62	Placebos/
63	placebo$.tw.
64	random$.tw.
65	research design/
66	or/58-65
67	66 not 56
68	comparative-study/
69	exp evaluation studies/
70	follow-up studies/
71	prospective studies/
72	(control$ or prospectiv$ or volunteer$).tw.
73	or/68-72
74	73 not 56
75	67 not 57
76	74 not (67 or 57)
77	28 and 57
78	28 and 75
79	28 and 76
80	or/77-79
81	33 and 57
82	33 and 75
83	33 and 76
84	or/81-83
85	39 and 57
86	39 and 75
87	39 and 76
88	or/85-87
89	44 and 57
90	44 and 75
91	44 and 76
92	or/89-91
93	from 92 keep 10

Table 5. Fibroids surgery search strategy

Table 5. Fibroids surgery search strategy

1	*uterine neoplasms/su
2	*uterine neoplasms/
3	*leiomyoma/su
4	*leiomyoma/
5	1 and 4
6	2 and 3
7	5 or 6
8	uterine fibroid$1.tw.
9	exp *hysterectomy/
10	myomectom$.tw.
11	9 or 10
12	8 and 11
13	2 and 4
14	13 and 11
15	7 or 12 or 14
16	limit 15 to (human and English language and yr = 1975-2000)
17	limit 16 to (clinical trial or clinical trial, phase i or clinical trial, phase ii or clinical trial, phase iii or clinical trial, phase iv or controlled clinical trial or meta analysis or multicenter study or review or review literature or review of reported cases or review, academic or review, multicase)
18	cohort.tw.
19	(prospective or retrospective).tw.
20	case series.tw.
21	exp cohort studies/
22	18 or 19 or 20 or 21
23	16 and 22
24	17 or 23
25	from 24 keep 1-166
26	from 25 keep 10

Table 6. Hysterectomy search strategy

Table 6. Hysterectomy search strategy

1	exp hysterectomy/
2	"myomectomy".mp.
3	hysterectom$.tw.
4	myomectom$.tw.
5	exp hysterectomy, vaginal/
6	"uterine artery embolization".mp.
7	or/1-6
8	exp ovary/
9	exp Ovariectomy/
10	Menopause/or Menopause, premature/
11	or/8-10
12	8 or 9
13	10 and 12
14	7 and 13
15	7 and 11
16	limit 15 to human
17	limit 15 to English
18	limit 16 to English
19	limit 18 to female
20	limit 19 to yr = 1975-2000
21	limit 14 to human
22	limit 21 to English
23	limit 22 to female
24	limit 23 to yr = 1975-2000
25	9 and 10
26	and 7
27	limit 26 to human
28	limit 27 to English
29	limit 28 to female
30	limit 29 to yr = 1975-2000
31	7 and 10
32	limit 31 to human
33	limit 32 to female
34	limit 33 to English
35	limit 34 to yr = 1975-2000
36	randomized controlled trials/
37	random allocation/
38	double-blind method/
39	Single-blind method/
40	randomized controlled trial.pt.
41	or/36-40
42	exp Cohort studies/
43	cohort.tw.
44	(prospective or retrospective).tw.
45	exp Clinical trials/
46	cross-sectional studies/
47	prospective studies/
48	Follow-up studies/
49	"comparative study"/
50	case series.tw.
51	or/42-50
52	41 or 51
53	20 and 52
54	23 and 52
55	24 and 52
56	30 and 52
57	35 and 52
58	exp sex disorders/
59	exp sex behavior/
60	58 or 59
61	and 7
62	and 52
63	exp Urination disorders/
64	exp Colonic diseases/
65	exp Fecal incontinence/
66	exp Uterine prolapse/
67	"VAGINAL PROLAPSE".mp.
68	exp Pelvic floor/
69	63 and 7
70	and 52
71	limit 70 to human
72	limit 71 to English language
73	limit 72 to female
74	limit 73 to yr = 1975-2000
75	64 and 65
76	64 or 65
77	and 7
78	and 52
79	limit 78 to human
80	limit 79 to English language
81	limit 80 to female
82	limit 81 to yr = 1975-2000
83	or/67-69
84	or/66-68
85	and 7
86	and 52
87	limit 86 to human
88	limit 87 to English language
89	limit 88 to female
90	limit 89 to yr = 1975-2000

Screening Criteria

Screening Results

The main searches (the basic search and the surgery search) yielded 1,084 English-language articles. The abstracts of these articles were reviewed against the inclusion/exclusion criteria by five physician investigators. A team of two physicians reviewed each abstract. Abstracts were available for more than three-fourths of the citations; however, when no abstract was available, the title, source, and MeSH words were reviewed. At this stage, articles were included if requested by one member of the review team.

Table 7. Results of kappa test

Table 7. Results of kappa test

	Reviewer 1	Reviewer 2	Reviewer 3	Reviewer 4
Reviewer 1	n/a	0.597	0.579	0.641
Reviewer 2	0.567	n/a	0.621	0.530
Reviewer 3	0.639	0.601	n/a	0.542
Reviewer 4	0.647	0.616	0.604	n/a

Notes: (1) The upper right numbers show kappa statistics based on dichotomizing the ratings to "include" versus "exclude." (2) The lower left numbers show kappa statistics based on multilevel ratings (1 through 8), representing the inclusion and exclusion criteria.

At the full-text screening stage, each article was independently reviewed by two physicians, who forwarded their decisions to the task order manager for recording and comparison. When necessary, the reviewers were asked to reconcile differences of opinion. Overall, the teams initially disagreed on about 20-25 percent of their decisions, but all disagreements were resolved.

The articles identified by the supplemental hysterectomy search were reviewed by a single reviewer with a special interest in gynecological surgery and the aging process. Eighty-eight of these articles were included in the discussion of Question 9; these articles were added to the project database.

Table 8. Number of records identified in main literature searches, (more...)

Table 8. Number of records identified in main literature searches, by database

Database	Number of records	Percent of total
MEDLINE	539	49.7
EMBASE	491	45.3
CancerLit	20	1.8
Cochrane	1	<1
PreMEDLINE	3	<1
HealthSTAR	2	<1
Other	28	2.6
Total	1,084	100

Table 9. Results of abstract screening and full-text (more...)

Table 9. Results of abstract screening and full-text article reviews

Description	No.	Percent
Abstracts
Number of abstracts reviewed (note: some records did not contain abstracts)	1,084	--
Number of relevant articles included:	637	58.8
1-Original research on treatment (for abstraction into evidence tables)	325	51.0
2-Review articles on treatment	141	22.1
3-Other (epidemiology, health services utilization)	76	11.9
4-Basic science	95	14.9
Number excluded:	447	41.2
Reasons for exclusion:
1-Not original research	37	8.3
2-Not uterine fibroids	38	8.5
3-Case report	138	30.9
4-Small case series (n < 20)	50	11.2
5-Background	184	41.1
Full-Text Articles
Number of full-text articles reviewed	637	--
Number of relevant articles:	550	86.3
1-Original research on treatment (for abstraction into evidence tables)	200	36.4
2-Review articles on treatment	144	26.2
3-Other (epidemiology, health services utilization)	95	17.3
4-Basic science	102	18.5
5-Other (no separate fibroids results)	9	1.6
Number excluded:	87	13.7
Reasons for exclusion:
1-Not original research	22	25.3
2-Not uterine fibroids	14	16.1
3-Case report	7	8.0
4-Small case series (n < 20)	40	46.0
5-Other (e.g., unobtainable, not English)	4	4.6

We also performed a review of Issue 3 of the 1999 Cochrane Database of Systematic Reviews. Fifteen citations for "fibroids" were found, three for "myomectomy," and 28 for "hysterectomy." However, only one review specifically addressed hysterectomy or myomectomy for fibroids, and the focus of this review was on preoperative treatment with GnRH agonists (Lethaby, Vollenhoven, and Sowter, 1999). Nineteen trials identified in this review met our inclusion criteria, and all 19 were identified in our online database searches. Two other Cochrane reviews focused on adjunctive treatments to prevent adhesions at the time of infertility surgery (including myomectomy). Although there were four Cochrane reviews of trials of medical therapies for heavy menstrual bleeding, all but one of the included trials explicitly excluded women with fibroids; this trial was identified in our online database search.

During the peer review process, additional articles were suggested by many of the reviewers. All suggested articles went through the same screening process as the original articles and were added to the evidence tables if they met our criteria. Several reviewers also pointed out case reports of particularly severe complications of certain procedures that were excluded by our screening criteria. In reviews of surgical procedures, excluding case reports or small case series clearly leads to omission of these reports. However, for the purposes of this review, we are more concerned with the question "How likely is something to happen?" than the question "Can something happen?" since almost any conceivable complication can occur with any procedure. In addition, it is likely that there is a bias toward both reporting small numbers of complications of newer procedures and publishing those results, since similar complications of more established procedures might be deemed less "reportable." Therefore, we have mentioned case reports of complications where more than one reviewer felt that discussion of these complications was important, but we did not revise our search and screening criteria or systematically review all case reports and small case series that had been previously excluded.

Internal Validity

The criteria for assessing internal validity were as follows:

Randomized allocation to treatment:

• Rationale: By randomly assigning groups to the intervention of interest, other factors that may confound the results are equally distributed between groups (assuming a large enough sample size). This equal distribution minimizes the chances of over- or underestimation of treatment effect based on unequal distribution of confounding factors.

• Operational definition: Criterion met if (and only if) the trial report explicitly stated that treatment was randomly allocated.

If randomized, appropriate randomization methods used:

• Rationale: "Pseudo-randomization" methods may be susceptible to bias, as demonstrated by evidence of unequal distribution of subject characteristics (Schulz, Chalmers, Grimes, et al., 1994) and larger effect sizes compared with studies using more rigorous methods (Schulz, Chalmers, Hayes, et al., 1995). In addition, methods of allocation concealment are also important in preventing bias (e.g., use of preprepared sealed envelopes).

• Operational definition: Criterion met if randomization methods were not susceptible to bias, such as computer-generated numbers in sealed envelopes. Criterion not met by studies that either used methods more prone to bias, such as alternate medical record numbers, or did not describe randomization methods or methods of allocation concealment.

Adequate description of patients and controls:

• Rationale: Patient characteristics that might affect outcomes (such as obesity, prior surgery, medical comorbidities) are likely to differ between two interventions. If these differences are not characterized, then erroneous conclusions may be drawn. For example, comparison of outcomes from a series of laparoscopic appendectomies with those from concurrent open appendectomies found better outcomes with the laparoscopic procedure (Sosa, Sleeman, McKenney, et al., 1993). These differences were not seen when the same group performed a randomized trial, a finding attributable to differential patient selection criteria in the nonrandomized study (Martin, Puente, Sosa, et al., 1995).

• Operational definition: Criterion met if (a) inclusion and exclusion criteria for participation in the study were described or (b) for nonrandomized studies, description of the rationale for selecting a particular intervention was given. Criterion not met if (a) inclusion/exclusion criteria were not described or (b) description of the rationale for selection of the interventions was not given (e.g., a nonrandomized comparison of concurrent laparoscopic and abdominal myomectomies that did not describe why patients received one or the other procedure).

Description of length of followup:

• Rationale: Patient and provider decision-making is dependent on knowledge of the likelihood of benefits and risks over time.

• Operational definition: Criterion met if a summary statistic (mean, median, or minimum) and range of followup were given for studies reporting nonhospital outcomes. Criterion not met if summary statistic and range not given. Criterion not applicable for studies reporting only hospital-based outcomes.

Description of loss to followup:

• Rationale: Failing to account for patients lost to followup may lead to erroneous conclusions, especially if the loss to followup is related to either the underlying disease or the intervention (e.g., patients seeking care elsewhere because of continuing symptoms or unacceptable side effects of treatment).

• Operational definition: Criterion met if (a) loss to followup was explicitly reported, (b) number of subjects for whom data was presented was equal to number of subjects receiving intervention at start of study, or (c) for studies reporting only hospital-based outcomes, number of missing charts or records was reported. Criterion not met if loss to followup was not reported and number of subjects at beginning and end of study was not equal.

Description of dropout rates:

• Rationale: Dropout rates may reflect differences in clinically important variables, such as side effects or treatment response. Failure to account for dropouts may result in erroneous conclusions similar to those seen with failure to account for loss to followup.

• Operational definition: Criterion met if (a) patients dropping out of the study prior to completion were reported or (b) number of subjects at beginning and end of study were equal. Criterion not met if patients dropping out were not reported and numbers of subjects at beginning and end of study were not equal. Criterion not applicable for studies reporting only hospital-based outcomes.

Recognition and description of statistical issues:

• Rationale: Use of inappropriate tests may lead to misleading conclusions. For example, variables such as blood loss, length of stay, or costs are often not normally distributed; use of means instead of medians when data may be affected by outlying observations can be misleading. Many studies, especially case series, may lack sufficient power to detect clinically important differences in outcomes or patient characteristics.

• Operational definition: Criterion met if (a) appropriate statistical tests were used (e.g., nonparametric methods for variables with nonnormal distributions, or survival analysis techniques to account for loss to followup and dropouts) and (b) potential study limitations regarding design and analysis, especially sample size and power issues, were discussed. Criterion not met if (a) inappropriate statistical tests were used or (b) study limitations were not discussed.

External Validity

The criteria for assessing external validity were as follows:

Description of age of study population:

• Rationale: The outcomes of many interventions are affected by patient age. Age is especially important in studies of reproductive disorders in women, since childbearing potential and ovarian hormone production, both key components in decisionmaking regarding management of fibroids, are directly related to age.

• Operational definition: Criterion met if summary statistics of subject age were given. Criterion not met if summary statistics were not given.

Description of racial/ethnic distribution of population:

• Rationale: The epidemiology, and possibly the biology, of fibroids clearly varies between white and black women. Additionally, there is widespread racial variation in the United States in utilization and outcomes of a wide variety of interventions (Fiscella, Franks, Gold, et al., 2000).

• Operational definition: Criterion met if (a) racial/ethnic distribution was described or (b) the geographical setting of the study strongly implied the racial/ethnic background of the entire population (e.g., studies of hysterectomy outcomes in Japan or Nigeria). Criterion not met if (a) racial/ethnic distribution was not described and (b) geographic setting was likely to include subjects of diverse racial/ethnic background.

Description of pregnancy history of population:

• Rationale: Pregnancy history may affect the natural history or biology of fibroids (Parazzini, Negri, La Vecchia, et al., 1996). For surgical interventions, pregnancy history may affect the technical difficulty of a procedure; for example, prior vaginal delivery may facilitate vaginal hysterectomy, while prior cesarean section, by increasing the risk of adhesions, may make either abdominal or vaginal hysterectomy more difficult.

• Operational definition: Criterion met if (a) summary statistics on gravidity or parity were given or (b) percentage of women with prior pregnancy was given. Criterion not met if (a) no summary statistics were given and (b) no distribution data on prior pregnancies were given.

Description of prior surgery:

• Rationale: A history of prior surgery for fibroids might reflect differences in the natural history or biology between patients. Additionally, previous abdominal surgery might increase the risk of complications by increasing the likelihood of intraperitoneal adhesions.

• Operational definition: Criterion met if (a) any description of history of intra-abdominal surgery was given or (b) proportion of women with prior surgery for fibroids was given. Criterion not met if no description of prior surgery was given.

Adequate characterization of fibroid and/or uterine size:

• Rationale: Individual fibroid size, or aggregate uterine size, may affect the nature or severity of symptoms, the response to various treatments, and the risk of complications of surgical treatments.

• Operational definition: Criterion met if data given on (a) uterine size in weeks gestational age; (b) uterine volume, area, or length as estimated by radiologic techniques; (c) uterine weight in grams (for hysterectomy specimens); (d) fibroid diameter or volume as estimated by radiologic techniques; or (e) fibroid dimensions or weight based on pathological examinations. Criterion not met if none of the above were provided.

Adequate characterization of fibroid number:

• Rationale: The number of fibroids may affect the nature or severity of symptoms, the response to various treatments, and the risk of complications of surgical treatments.

• Operational definition: Criterion met if summary statistics or distribution of number of fibroids was provided. Criterion not met if no data were provided on number of fibroids.

Adequate characterization of fibroid location:

• Rationale: The location of fibroids may affect the nature or severity of symptoms, the response to various treatments, and the risk of complications of surgical treatments.

• Operational definition: Criterion met if (a) distribution of fibroids by location (subserosal, intramural, submucosal, or pedunculated) was given or (b) other anatomical descriptions were given (e.g., anterior, posterior, fundal, or within the broad ligament). Criterion not met if no anatomical description was given.

Adequate characterization of baseline symptoms:

• Rationale: Because fibroids may present with a variety of symptoms, assessing the effectiveness of therapy requires an adequate description of the nature and severity of symptoms prior to institution of therapy.

• Operational definition: Criterion met if distribution of specific symptoms or symptom classes associated with fibroids were provided. Criterion not met if specific symptoms were not described (e.g., if the only description of inclusion criteria was "symptomatic fibroids").

Adequate description of timing of outcome measurement:

• Rationale: Outcome measures may vary depending on when they are obtained. Description of when outcomes were measured facilitates comparison between studies.

• Operational definition: Criterion met if (a) time after initiation of therapy at which outcomes were measured was reported or (b) study was limited to hospital-based outcomes. Criterion is not met if (a) time was not reported and (b) study was not strictly hospital-based.

Adequate description of methods used for outcome measurement:

• Rationale: Comparison between studies requires common methods of measurement, which in turn requires adequate description of the methods used to assess comparability.

• Operational definition: Criterion met if (a) methods used to measure outcomes were adequately described or referenced (e.g., pain or bleeding scales), (b) definitions were given (e.g., description of outcomes classified as "complications"), or (c) outcomes were unambiguous (e.g., pregnancy, need for hysterectomy). Criterion not met if (a), (b), or (c) was not present.

Adequate description of validity and reliability of outcome measurement:

• Rationale: Measurements of outcomes are only useful if changes in the outcome being measured are reflected in changes in the measurement (validity) and if these changes are reasonably consistent between the same observer measuring at different times or between different observers (reliability). For example, changes in a scale to assess menstrual blood flow should correlate with some other physiological measure of menstrual blood loss, and this correlation should be consistent when different women apply the same scale.

• Operational definition: Criterion met if (a) a description of the methods used to assess validity and reliability of at least one outcome measure was provided, (b) a reference to another article documenting validity and reliability was provided, or (c) only unambiguous outcomes such as pregnancy were included. Criterion not met if (a), (b), or (c) was not present.

Adequate description of clinical care provided to subjects:

• Rationale: The ability to replicate study results is dependent on adequate description of methods. Additionally, readers should be aware of aspects of clinical care that might influence outcomes.

• Operational definition: Criterion met if (a) a detailed description of the therapy (dose, dosing schedule, and route of administration for medications and/or techniques for invasive therapies) was provided; (b) a reference to another publication describing the procedure was provided; or (c) statistical adjustment was made for likely sources of variation in clinical care (e.g., site where care was given, type of specialist providing care, individual provider). Criterion not met if (a), (b), or (c) was not provided.

Use of previously validated and standardized measures:

• Rationale: Use of measures used by other researchers enhances the ability to compare results across studies. Use of measures used with other medical conditions enhances the ability to compare the impact of uterine fibroids to that of other common conditions, which may be important when setting research and resource allocation priorities.

• Operational definition: Criterion met if at least one measure previously used by another group was used. Criterion not met if all measures were internally developed.

Reporting of Quality Scores

These criteria were not aggregated into an overall quality score but were considered and reported individually. This approach was preferred for several reasons:

• Previous work has shown that numeric grading systems may not discriminate well between high- and low-quality studies, even randomized trials (Juni, Witschi, Bloch, et al., 1999; Moher, Jadad, and Tugwell, 1996).

• Development and use of a new quality score would require additional work for validation.

• Identification of specific weaknesses in each study will be helpful in identifying trends, which in turn will assist with recommendations for future research.

Our approach of describing key study design components, rather than assigning a single aggregate score, is also consistent with recent recommendations from an expert panel on meta-analysis of observational studies (Stroup, Berlin, Morton, et al., 2000).

Summaries of each evaluation are provided in the evidence table entry for each abstracted article. A "+" indicates that a given criterion was met, a "−" denotes that the criterion was not met. Grades were assigned by the primary reviewer and confirmed by the over-reader.

Nationwide Inpatient Sample

The NIS contains administrative discharge data from more than 1000 hospitals in 22 States, representing a stratified sample of 20 percent of U.S. hospitals (NIS, 1997). Weights are provided to permit estimation of national data based on this sample. We used this data set to provide supplemental data on frequency and resource utilization for hysterectomy and myomectomy by age and race (Question 7 on cost and Question 8 on age and racial variation). Because previous work has shown that administrative data may lack sufficient clinical detail to compare outcomes (Myers and Steege, 1999), we did not attempt to directly compare complication rates between these procedures (Question 1).

Hysterectomies for fibroids were identified by searching the 1997 NIS data set for all patients with International Classification of Diseases, Revision 9 (ICD-9) procedure codes of 68.3, 68.4, and 68.5 (all nonradical hysterectomies) who also had diagnosis codes of 218.0, 218.1, 218.2, and 218.9 (submucous, intramural, subserous, and unspecified fibroids).

Myomectomies were identified by searching the 1997 NIS data set for all patients with ICD-9 procedure codes of 68.29 (uterine lesion destruction) who also had diagnosis codes of 218.0, 218.1, 218.2, and 218.9. We used these codes to identify myomectomies at Duke after discussion with hospital coders; only three nonmyomectomy cases were identified using this strategy. Further validation of this coding strategy, especially trying to estimate the proportion of myomectomies not identified, would be useful for future health services research.

Data from the sampled hospitals were converted to national estimates using the weighting variables provided by AHRQ. To assess differences in the proportion of white and black women undergoing each procedure, logistic regression was performed using myomectomy versus hysterectomy as the dependent variable, race as the main independent variable, and controlling for age, payer type, and median income.

Duke University Medical Center Data

Data on hospital costs and clinical characteristics affecting cost and complications of patients undergoing myomectomy were obtained from Duke University Medical Center. Subsequent analyses will include additional chart data on patients undergoing hysterectomy for fibroids. This data source was used to provide additional information for Question 1 (risks and benefits of hysterectomy and myomectomy), Question 7 (costs of interventions), and Question 8 (racial differences in outcomes).

Meta-analysis

Decision Analysis

Decision-analytic models can be a useful means of synthesizing information from a variety of sources, and they have been used by the Duke group in studies of stroke (Matchar, Samsa, Matthews, et al., 1997) and cervical cytology (McCrory, Matchar, Bastian, et al., 1998). We had planned on using a decision model to synthesize the disparate data on fibroid management to help answer the key questions or at least to identify key areas of uncertainty to direct future research. We were able to use decision modeling for some purposes (e.g., estimating cumulative incidence of fibroids based on reported age-specific rates). However, the methodological limitations of the available literature precluded us from using decision analysis as a formal synthesis tool. In particular, heterogeneity of methods for measuring and reporting patient outcomes and for reporting clinically important preintervention patient characteristics limited our ability to perform even preliminary decision analyses.

Developing a decision model in parallel with our literature review did enable us to consider the nature of the data and the potentially important parameters that would be needed to use such a model to assist patients, clinicians, or policymakers. Therefore, we have provided a detailed description of the framework for such a model, along with some potential examples of its use, in Chapter 4.

Evidence Tables

The evidence tables summarize each study meeting the inclusion criteria described above; as such, they represent studies reporting on outcomes of prophylactic or symptomatic treatments for fibroids. Articles cited in the text but not in the evidence tables include (1) review articles; (2) articles addressing background issues such as basic biology, epidemiology, or health services research; and (3) articles either describing methods or illustrating relevant methodological issues.

Many of the studies included in the review of the effects of hysterectomy on the aging process (Question 9) also were not abstracted in evidence tables. Because our initial searches focused on patients with fibroids, many of the articles discussed under Question 9 did not meet our original inclusion criteria. Late in the review process, after it became apparent that the vast majority of studies addressing this question would be excluded, we revisited the question and performed a revised search. Because a large proportion of women, especially premenopausal women, who undergo hysterectomy do so because of symptomatic fibroids, we assumed that these articles would have some validity for women with fibroids. As mentioned above, the review of these articles did not follow the same strict methodology used in the review of articles from the basic search and the surgery search. For that reason, and because they did not explicitly focus on women with fibroids, these articles were not abstracted into the evidence tables.

Because formal data synthesis techniques were not applicable to the majority of the literature on management of fibroids, we elected to summarize the results of our review without attempting quantification. We did attempt to identify qualitatively similar findings that could be suggestive of consistent results. We also tried to identify the characteristics of data that would be needed to address the question adequately and, in some cases, to identify conceptual issues that should be considered by patients and providers given the lack of data (issues surrounding prophylactic surgery for asymptomatic fibroids, for example).

Approach

In the absence of randomized trials comparing hysterectomy and myomectomy in women with symptomatic fibroids, we reviewed the literature on myomectomy and hysterectomy outcomes separately, as well as the few studies that compared nonrandomized patients undergoing the two procedures.

We did not identify any studies that specifically described outcomes in women undergoing surgery for asymptomatic fibroids. However, because many women may be advised to undergo such surgery -- 11 percent of all of the hysterectomies done for fibroids in the Maine Women's Health Study were because of physician concern about delaying treatment (Carlson, Miller, and Fowler, 1994a) -- we reviewed the potential arguments for and against such surgery and examined the literature on prophylactic myomectomy and hysterectomy (see Evidence Table 1).

Myomectomy for Symptomatic Fibroids

Table 10. Myomectomy for symptomatic fibroids

Table 10. Myomectomy for symptomatic fibroids

Study	No. of patients	Length of followup	Relief of symptoms	Short-term complications	Long-term complications
Laparoscopic myomectomy
Andrei, Crovini, and Rosi, 1999	332	Not reported	Not reported	Transfusion: 1.2%	Not reported
Carter and McCarus, 1997	28	6 months	Bleeding: 100% "satisfied with current bleeding pattern" Pain: 80% "whose chief complaint was pain were satisfied with current levels of pain"	Not reported	Not reported
Chapman, 1998	300 (laser-induced interstitial thermo-therapy)	6-72 months	Bleeding: 89% (n = 276) Pain: 85% (n = 156) Dyspareunia: 100% (n = 4) Urinary sx: 100% (n = 24)	Not reported	Not reported
Daniell, 1995	32	6-36 months	Bleeding: 100% (n = 10) Dysmenorrhea: 60% full, 40% partial relief (n = 10) Pelvic pain: 29% full, 54% partial, 14% no relief (n = 7)	Not reported	Not reported
Darai, Dechaud, Benifla, et al., 1997	143	Minimum 12 months	Not reported	Transfusion: 1.1%	Not reported
Hasson, Rotman, Rana, et al., 1992	56	Mean 9 months	Not reported	Transfusion: 0% Subcutaneous emphysema: 2%	Not reported
Nezhat, Nezhat, Bess, et al., 1994	57	Mean 11 months	Menorrhagia: 100% complete relief Pelvic pain: 14/23 complete, 9/23 partial relief	1 hernia, 1 pneumonia	Not reported
Nezhat, Nezhat, Silfen, et al., 1991	158	3 months-3 years	Not reported	2.5% major complications	Not reported
Nezhat, Roemisch, Nezhat, et al., 1998	114	Mean 37 months	Not reported	Transfusions: 7%	Not reported
Rossetti, Paccosi, Sizzi, et al., 1999	123	Not reported	Not reported	0	Not reported
Seinera, Arisio, Decko, et al., 1997	54	6 months	Bleeding (n = 20): 90% had "no complaints"	Hemorrhage requiring reoperation: 1.8%	Not reported
Stringer, 1996	41	Mean 18 months	91% "complete or significant relief of symptoms" (type not specified)	Not reported	Not reported
Hysteroscopic myomectomy
Baggish, Sze, and Morgan, 1989	23	Not reported	90% "normal menses"	Pulmonary edema: 4%	Not reported
Bernard, Darai, Poncelet, et al., 2000	31	Minimum 12 months	Not reported	Uterine perforation: 3.3% Hemorrhage: 3.3%	Not reported
Brooks, Loffer, and Serden, 1989	62	Minimum 3 months	Menorrhagia: 46/57 "improved," 4/57 "not improved"	Not reported	Not reported
Corson and Brooks, 1991	92	Mean 17 months	Menorrhagia: "Success" 81% Pain: "Success" 85%	Transfusion: 1.1% Uterine perforation: 3.2% Fever: 1.1%	Not reported
Cravello, D'Ercole, Boubli, et al., 1995	239	Mean 2.5 years	Bleeding (premenopausal): 81% "satisfied and could lead normal life"	Uterine perforation: 2.3%	Not reported
Cravello, Farnarier, De Montgolfier, et al., 1999	196	Mean 73 months	68.4% "improved"	Not reported	Not reported
De Blok, Dijkman, and Hemrika, 1996	109	Mean 2.8 years	93% "improved"	Not reported	Not reported
Derman, Rehnstrom, and Neuwirth, 1991	108	1-16 years	75% without recurrent bleeding	Transfusions: 3.7%	Not reported
Donnez, Gillerot, Bourgonjon, et al., 1990	60	Not reported	100% "normal flow"	0%	Not reported
Dubuisson, Chapron, and Levy, 1996	213	Not reported	Not reported	3.8%, significantly related to increasing fibroid size	Not reported
Dueholm, Forman, and Ingerslev, 1998	45	Minimum 3 months	Not reported	Uterine perforation: 4.4%	Not reported
Emanuel, Wamsteker, Hart, et al., 1999	285	16-104 months	Not reported	Uterine perforation: 0.3% Pulmonary edema: 0.3%	Not reported
Giatras, Berkeley, Noyes, et al., 1999	41	Not reported	Bleeding: 80% "resolved"	Not reported	Asherman's syndrome: 1/41
Hallez, 1995	274	5-63 months	67% free of symptoms after 60 months	Uterine perforation: 0.3%	(At 6 months) Bleeding: 4% Asherman's: 10%
Hart, Molnar, and Magos, 1999	122	Mean 2.3 years	Menstrual bleeding 86% "significant improvement" Dysmenorrhea: 72% "significant improvement"	Not reported	11% uterine rupture rate during subsequent pregnancy
Indman, 1993	38 with concurrent ablation, 13 without	Mean 2.2 years	With ablation: 92% extremely satisfied Without ablation: 60% extremely satisfied	Not reported	Not reported
Itzkowic, 1993	40	Mean 14 months	Bleeding: 83% improved	Infection: 2.5%	Not reported
Lin, Iwata, and Liu, 1994	25 (pre-treated with GnRH)	9 months	Menorrhagia: 88% improvement	Cervical laceration: 4%	Not reported
Loffer, 1990	43	Minimum 12 months	Excessive menstrual bleeding: "controlled in 93%"	Perforation: 1.3% Hyponatremia: 3.8%	Not reported
Lomano, 1991	33	Mean 8 months	88% with heavy/severe menstrual bleeding pre-op; 0% post-op	Not reported	Not reported
Mints, Radestad, and Rylander, 1998	62	Mean 29 months	74% of patients with menorrhagia had post-op amenorrhea/hypo-menorrhea	Not reported	Not reported
Motashaw, Dave, and Paghdiwalla, 1995	73	Not reported	83% premenopausal patients "happy with result"	1% emergent hysterectomy Transfusions: 4% Perforation: 1%	Not reported
Phillips, Nathanson, Meltzer, et al., 1995	208 (120 with myomectomy alone, 88 with myomectomy + endometrial ablation)	6 months (n = 208) 6 years (n = 185)	Resection only: 85% with "satisfactory" results Resection + ablation: 88.5% with "satisfactory" results	Not reported	Not reported
Abdominal myomectomy
Berkeley, DeCherney, and Polan, 1983	50	Median 50.7 months	Not reported	Transfusion: 18% Fever: 28% Reoperation: 2%	Not reported
Brown, Fletcher, Myrie, et al., 1999	16 (done at time of c-section)	Not reported	Not reported	Transfusion: 1/16 (not different than c-section without myomectomy)	Not reported
Egwuatu, 1989	55	1-4 years	Not reported	Fever: 71.8% Wound infection: 10.9% Hemorrhage: 2.7%	Not reported
Ikpeze and Nwosu, 1998	72	Not reported	Not reported	Transfusion: 15.3% Fever: 17%	Not reported
LaMorte, Lalwani, and Diamond, 1993	128	Not reported	Not reported	Transfusion: 20% Fever: 12% Hysterectomy at time of procedure: 1%	Not reported
Reilly and Nour, 1998	120	Not reported	Not reported	Autologous transfusions: 4.2% Fever 11.6%	Not reported
Sirjusingh, Bassaw, and Roopnarinesingh, 1994	114	2-10 years	Not reported	Transfusion: 16%	Not reported
Vercellini, Maddalena, De Giorgi, et al., 1999	466	41 months	Not reported	Transfusion: 2% Fever: 18% Reoperation: 2%	Not reported
Vaginal myomectomy
Ben-Baruch, Schiff, Menashe, et al., 1988	46 (all solitary prolapsed, pedun-culated, sub-mucous)	Median 5.5 years	Asymptomatic: 79%	Transfusion: 8.7% pre-op conversion to abdominal procedure: 6.5%	Not reported
Davies, Hart, and Magos, 1999	35	Not reported	Menorrhagia: 15/21 "improvement" Pelvic pain: 6/7 "improvement"	Transfusion: 11.4% Hematoma: 11.4% Reoperation or readmission: 8.5%	Not reported

Transfusion was the most commonly reported short-term complication, with rates varying from 1.2 percent to 18 percent. However, wide variations in practice setting and local practices, as well as time and the availability of autologous blood, probably influenced these rates. Uterine perforation and fluid/electrolyte disturbances are a unique risk with hysteroscopic procedures; again, lack of comparability between studies precluded quantitative estimation of this risk.

Only three studies, all of hysteroscopic procedures, reported long-term adverse outcomes. Two reported the development of intrauterine scars (Asherman's syndrome) in up to 10 percent of patients. This may not be a concern for women not planning future pregnancy, but it would have an adverse effect on women's ability to conceive.

Hysterectomy for Symptomatic Fibroids

Table 11. Hysterectomy for symptomatic fibroids

Table 11. Hysterectomy for symptomatic fibroids

Study	No. of patients	Length of followup	Relief of symptoms	Short-term complications	Long-term complications
Carlson, Miller, and Fowler, 1994	413 (35% for fibroids)	12 months	Significant improvement in symptoms in all women with fibroids	In-hospital complications: 7%	New symptoms: Hot flashes: 13% Weight gain: 12% Depression: 8% Anxiety: 7%
Kjerulff, Langenberg, Rhodes, et al., 2000	1299 (48.1% for fibroids)	24 months	Results not stratified by indication (not significantly different by indication, according to authors) All symptoms significantly improved by hysterectomy; quality of life also improved	None: 21.4% Mild: 66.8% Moderate: 11.1% Severe: 0.7%	Problems acquired in up to 12.9% Predictors of poor long-term outcome: Income, baseline depression/anxiety, and bilateral salpingoophorectomy
Weber, Walters, Schover, et al., 1999	43 (71% for fibroids)	Mean 14.2 ± 3.5 months	Most important benefit of surgery: Relief from pain: 55% Relief from bleeding: 40% 95% satisfied	None: 100%	4/34 without incontinence pre-operatively developed stress incontinence post-operatively

Hysterectomy Versus Myomectomy for Symptomatic Fibroids

Table 12. Studies directly comparing abdominal hysterectomy (more...)

Table 12. Studies directly comparing abdominal hysterectomy to abdominal myomectomy

Reference	No. of patients	Short-term complications	Long-term complications	Comments
Ecker, Foster, and Friedman, 1995	204 abdominal hysterectomy; 109 abdominal myomectomy	1) Mean estimated blood loss (EBL) greater in hysterectomy group 2) EBL correlated with increasing uterine size, number of fibroids removed, and length of procedure 3) Febrile morbidity more common in myomectomy group 4) Myomectomy group more likely to have banked blood, but no difference in transfusion rates	Not reported	Retrospective chart review No multivariate analysis
Iverson, Chelmow, Strohbehn, et al., 1996	89 abdominal hysterectomy; 103 abdominal myomectomy	1) EBL greater in hysterectomy group (univariate) 2) Uterine size only significant predictor of EBL in multivariate analysis 3) Increased intraoperative visceral injuries in hysterectomy group (no multivariate analysis)	Not reported	Retrospective chart review
Iverson, Chelmow, Strohbehn, et al., 1999	160 abdominal hysterectomy; 101 abdominal myomectomy	Myomectomy associated with significantly increased risk of fever compared with hysterectomy (multivariate analysis)	Not reported	None

Hysterectomy or Myomectomy for Asymptomatic Fibroids

We were unable to identify any studies that provided direct evidence on risks or benefits of surgery for women with asymptomatic fibroids. Historic arguments for performing surgery on women with asymptomatic fibroids above a certain size have included the following:

• Prevention of ovarian cancer mortality: Decreased sensitivity of bimanual pelvic examination for detecting adnexal pathology (especially ovarian malignancy) secondary to the large uterus.

• Prevention of uterine cancer mortality: An increased risk of uterine sarcoma in large or rapidly growing fibroids.

• Prevention of fibroid growth after menopause: An increased risk of continued fibroid growth in women using postmenopausal hormone replacement therapy.

• Prevention of injury to other organ systems: An increased risk of compromise of adjacent organs (e.g., ureteral obstruction) secondary to compression by an enlarging uterus.

• Prevention of problems related to infertility or pregnancy complications: An increased risk of problems with either fertility or pregnancy with increasing fibroid size (valid only for myomectomy).

• Prevention of future surgical morbidity: An increased perioperative morbidity associated with interval growth of the uterus (Friedman and Haas, 1993).

Table 13. Studies evaluating the effect of uterine (more...)

Table 13. Studies evaluating the effect of uterine size on hysterectomy complications

Study	No. of patients	Length of followup	Relief of symptoms	Short-term complications	Long-term complications
Hillis, Marchbanks, and Peterson, 1996	446	6 weeks	Not reported	EBL, transfusions, and complications increased for uterine size > 500 g (multivariate analysis)	Not reported
Reiter, Wagner, and Gambone, 1992	93	Not reported	Not reported	Estimated blood loss (EBL) and complication rates not significantly different between uterus < 12 weeks size and > 12 weeks size	Not reported

• Prevention of ovarian cancer mortality: There are no data supporting the utility of a bimanual examination in women with normal-sized uteri as a screening tool for reducing ovarian cancer mortality (Grover and Quinn, 1995; Schutter, Kenemans, Sohn, et al., 1994). Therefore, the inability to palpate the ovaries in a woman with an enlarged uterus seems unlikely to have a substantial impact on ovarian cancer mortality.

• Prevention of uterine cancer mortality: Uterine leiomyosarcoma is rare, with a reported incidence of fewer than 1 per 100,000 women (Van Dinh and Woodruff, 1982). Sarcomas were found in 0.5 percent or less of all hysterectomies performed for suspected fibroids in most series (Leibsohn, d'Ablaing, Mishell, et al., 1990; Parker, Fu, and Berek, 1994; Takamizawa, Minakami, Usui, et al., 1999). Even in women for whom the preoperative diagnosis was "rapidly enlarging uterus," the prevalence was 0.27 percent (Parker, Fu, and Berek, 1994). Given the prevalence of fibroids and surgical procedures for fibroids and the low incidence of uterine sarcomas it seems likely that many sarcomas are discovered by serendipity (McNaughton Collins, Ransohoff, and Barry, 1997; Ransohoff and Lang, 1990).

• Prevention of growth after menopause: Although submucous fibroids may increase the risk of abnormal bleeding in women on hormone replacement therapy (Akkad, Habiba, Ismail, et al., 1995), data on growth in women on hormone replacement are limited and suggest that growth may be variable and depend on the regimen used (Sener, Seckin, Ozmen, et al., 1996; Ylostalo, Granberg, Backstrom, et al., 1996). If growth alone is not a compelling reason for surgery in premenopausal women, there are no obvious reasons why asymptomatic growth after menopause would be a reason for surgery.

• Prevention of injury to other organ systems: An enlarged uterus can result in distortions of urinary tract anatomy, including partial or complete ureteral obstruction (Piscitelli, Simel, and Addison, 1987). This may be exacerbated during pregnancy, when partial hydronephrosis is normal and, in rare cases, may result in acute renal injury (Courban, Blank, Harris, et al., 1997). However, there is no evidence to allow estimation of the risk of permanent urinary tract or renal injury associated with asymptomatic uterine growth. In fact, given a reported risk of urinary tract injury at hysterectomy of between 0.1 and 1.5 percent (Aslan, Brooks, Drummond, et al., 1999; Courban, Blank, Harris, et al., 1997; Daly and Higgins, 1988; Harkki-Siren, Sjoberg, and Tiitinen, 1998), the risk of permanent ureteral or renal injury secondary to uterine growth prior to the development of other symptoms would have to be at least as large to justify surgery on the basis of preventing obstructive organ injury.

• Prevention of problems related to infertility or pregnancy complications: Clearly, hysterectomy is inappropriate for any woman wishing to retain the option of future childbearing. The effects of fibroids on women seeking to conceive and on pregnancy outcomes are discussed under Question 8. However, there are no data on the impact of prophylactic myomectomy in asymptomatic women on either fertility or pregnancy outcomes. Given the frequency with which adhesions are found after myomectomy (March, Boyers, Franklin, et al., 1993; Tulandi, Murray, and Guralnick, 1993), it is possible that prophylactic myomectomy would have a harmful effect on fertility.

• Prevention of future operative morbidity: Two studies have specifically addressed the risk of perioperative morbidity and uterine size (Table 13). Reiter, Wagner, and Gambone (1992) reviewed the records of 93 women undergoing hysterectomy for fibroids at the San Diego Naval Medical Center. They found no statistical difference in estimated blood loss or complication rates between women with estimated uterine sizes less than 12 weeks gestational age (about 500 grams) and women with larger uteri. Conversely, Hillis, Marchbanks, and Peterson (1996) reviewed the records of 446 women who were enrolled in a prospective study of women undergoing sterilization procedures. Through both univariate and multivariate analyses they found that morbidity, especially transfusion risk, was increased in women with a uterine weight greater than 500 grams. Given the larger sample size, the prospective collection of data, and the use of more appropriate statistical methods, the findings of Hillis and colleagues (1996) are somewhat stronger.

Other indirect evidence supports an increased risk of perioperative morbidity with increasing uterine size. Ecker and colleagues (Ecker, Foster, and Friedman, 1995) and Iverson and colleagues (Iverson, Chelmow, Strohbehn, et al., 1996) both reported a statistically significant increase in estimated blood loss with increasing uterine size in univariate and multivariate analyses. Data from randomized trials of preoperative GnRH agonists show consistently decreased blood loss and operative times with preoperative GnRH agonists, which may be attributable to decreased uterine size (Lethaby, Vollenhoven, and Sowter, 1999).

Although the evidence suggests that increased uterine size is associated with increased blood loss and possibly an increased risk of other complications, this association does not necessarily lead to the conclusion that surgery on women with fibroids above a certain size will reduce morbidity for several reasons:

• None of the studies specifically reported on outcomes in women with asymptomatic fibroids compared with outcomes in women with symptomatic fibroids. It is possible that the natural history of asymptomatic fibroids -- including the rate of growth, number, and location within the uterus -- is different from that of symptomatic fibroids, and this may result in different perioperative outcomes.

• Use of preoperative GnRH agonists may reduce some risks associated with interval increase in uterine size (Lethaby, Vollenhoven, and Sowter, 1999).

• Without good data on the natural history of fibroids, it is difficult to estimate the likelihood that a woman with asymptomatic fibroids will either develop symptoms or reach an arbitrary size threshold within a given time period.

• Using estimates of a substantially increased relative risk of perioperative morbidity associated with increasing fibroid size will overestimate the benefits and underestimate the risks of immediate surgery if the probability of progression is not considered. For example, if the complication rate for a hysterectomy performed with a uterine size of 10 weeks is 10 percent, a two-fold increase in risk would increase the rate to 20 percent. For immediate surgery to be the preferred strategy in terms of reducing morbidity, the probability that the uterus would grow to a size that would result in twice the risk in the absence of any symptoms would need to be greater than 50 percent. If the probability of progression were 40 percent, then the overall risk of complications would be 0.4 x 0.2, or 8 percent, less than the 10 percent risk with immediate surgery.

Approach

This is a somewhat ambiguous question, which can be interpreted in two ways:

• What are the risks associated with a primary myomectomy compared with the risks associated with repeat myomectomies?

• What are the risks of a myomectomy that results in the removal of a single fibroid compared with the risks associated with the removal of multiple fibroids?

We clarified this through discussions with the advisory panel. Because Question 5 ("Does additional treatment result in significantly increased morbidity?") addresses the first interpretation, we focused on the second issue. We also determined that the question referred to the comparison between women with a single clinically detectable fibroid and women with multiple fibroids, since it would be highly unlikely that any surgeon would not remove all detectable fibroids. In other words, we assumed that once a decision to operate had been made, no surgeon would deliberately leave additional detectable fibroids in situ unless there were compelling intraoperative events (e.g., significant blood loss or complications with anesthesia that required shortening surgical time). In addition to examining the relevant literature, we analyzed a primary data set that allowed us to compare complication rates for women undergoing removal of a single fibroid with those for women undergoing removal of multiple fibroids. The outcomes we considered were perioperative complications, pregnancy rates, recurrence rates, and need for subsequent surgery.

Literature Search

Table 14. Studies comparing single to multiple myomectomies (more...)

Table 14. Studies comparing single to multiple myomectomies

Study	No. of patients	Length of followup	Complications	Pregnancy rate	Recurrence (diagnostic or symptomatic)	Need for subsequent surgery
Acien and Quereda, 1996	80	10 years	Not reported	1 fibroid: 79.2% > 2: 37.5% (p < 0.05)	Not reported by number of fibroids	Not reported by number of fibroids
Candiani, Fedele, Parazzini, et al., 1991	622	10 years	Not reported	Not reported	Multivariate relative risk (with 95% confidence interval): 1 fibroid: 1.0 2-3 fibroids: 1.2 (0.9-1.51) > 4 fibroids: 2.1 (1.7-2.8)	Not reported
Fedele, Parazzini, Luchini, et al., 1995	145	5 years	Not reported	Not reported	Single fibroid: 38% (ultrasound) Multiple: 60%	Not reported
Malone, 1969	125	Minimum 5 years	Not reported	1 fibroid: 59% > 2: 48%	Single: 26.6% Multiple: 58.8%	Single: 11.1% (all hysterectomies) Multiple: 26.3% (22.5% hysterectomies)

Primary Data

Data from the Duke primary chart abstraction suggest a relationship between the number of fibroids removed and complication risk, but the threshold number of fibroids may be somewhat greater than one. For patients with a single fibroid removed (n = 43), the complication rate was 20.9 percent; for those with two to four fibroids removed (n = 45), the complication rate was 11.1 percent; and for those with five or more fibroids removed (n = 112), the complication rate was 42 percent; this trend was highly significant (p < 0.0001). Unfortunately, data on fibroid location were not provided in the charts or operative notes. One difficulty with retrospective studies in teaching hospital settings is that the potential effects of surgeon experience cannot readily be determined; it is often impossible to discern who performed which portion of a given procedure from an operative report.

In subsequent multivariate analyses, using fibroid number as a continuous variable, increasing the number of fibroids removed was a significant predictor of complications and transfusions. Given the relatively small sample size, we were unable to determine a "threshold" number of fibroids at which risk becomes "unacceptable." However, it seems likely that, given a favorable location, removal of two adjacent fibroids would not result in substantially greater morbidity compared with removal of a single fibroid. Other factors besides the number of fibroids removed -- such as size, location, degree of vascularization, ease of identification of surgical planes -- are likely to play a role in determining surgical difficulty.

Approach

Although this question was originally intended to focus only on the issue of myomectomy versus hysterectomy, discussions with members of the project's advisory panel led to an expansion to include the other strategies (including no intervention along with other medical and surgical therapies).

The concept of "appropriateness" can have multiple meanings. One frequently used definition is that a procedure is appropriate if a patient undergoing the procedure received all diagnostic evaluations and alternative treatments recommended by a panel of experts (Broder, Kanouse, Mittman, et al., 2000). Alternatively, the most appropriate therapy can be defined as the procedure, treatment, or treatment strategy most likely to result in favorable short- and long-term outcomes, given relevant clinical factors and patient preferences. We discuss the implications of the two definitions below:

• Appropriateness criteria: Broder and colleagues found that 70 percent of hysterectomy cases in Southern California over a 2-year period did not meet the standards of expert panel recommendations, and 76 percent did not meet criteria defined by ACOG (Broder, Kanouse, Mittman, et al., 2000). In devising the RAND appropriateness criteria (Shekelle, Kahan, Bernstein, et al., 1998), the expert panel was hampered by the same lack of high-quality evidence that the we faced in writing this evidence report. Given the lack of evidence supporting many of the criteria in both the expert panel and ACOG recommendations, it is somewhat difficult to conclude that ". . . the care leading to recommendations of hysterectomies in our cohort was suboptimal" (Broder, Kanouse, Mittman, et al., 2000). For example, 9 percent of all cases were "premenopausal women with leiomyomata, bleeding, and pain who ha[d] significant anemia or functional impairment but ha[d] not undergone endometrial sampling," and all of these were judged inappropriate. However, given (a) the low incidence of endometrial carcinoma in premenopausal women; (b) the likelihood that many women with symptomatic fibroids will have undergone ultrasound evaluation, which by itself has acceptable sensitivity and specificity for detecting benign and malignant endometrial changes associated with abnormal bleeding (Tahir, Bigrigg, Browning, et al., 1999); (c) the fact that the presence of fibroids may decrease the ability to obtain adequate tissue on endometrial sampling (Gordon and Westgate, 1999); and (d) the lack of evidence that failure to sample the endometrium in a woman with fibroids associated with bleeding and pain scheduled to undergo hysterectomy would lead to a significant change in outcomes, the "requirement" for endometrial sampling in all women with this indication may be questioned, at least on cost-effectiveness grounds (Dunn, Stamm, Delorit, et al., 2000; Farquhar, Lethaby, Sowter, et al., 1999). Similarly, given the paucity of evidence supporting the effectiveness of medical treatment for symptomatic fibroids, requiring documentation of a trial of such treatment prior to proceeding with hysterectomy as a marker of "optimal" care for all patients seems difficult to justify.

• Choosing the treatment most likely to result in benefit to the patient: This interpretation of "appropriateness" is more clinically relevant and more consistent with the overall purpose of this report. Answering this question depends on the availability of evidence that allows clinicians and patients to estimate the likelihood of benefits and harm over both the short and long term for each procedure or treatment given defined symptoms, pathology, and comorbidities. Unfortunately, the available data do not allow us to answer this question for the majority of fibroid management strategies, using either direct evidence from clinical studies or decision analytic methods (see Question 4), a difficulty faced by other groups as well (Broder, Kanouse, Mittman, et al., 2000). There are some scenarios where preliminary conclusions can be drawn, based either on consistency of evidence, a lack of evidence supporting alternative strategies, or common sense (e.g., hysterectomy is clearly inappropriate for women planning future pregnancies). The discussion of each strategy below summarizes more detailed discussions of the evidence in other sections of this report.

Results

These are discussed in the appropriate sections for each individual treatment option.

Methodological Issues

These are discussed in the appropriate sections for each individual treatment option.

Summary

Approach

We approached this question by dividing it into four separate clinical scenarios:

• Initial treatment with medical therapy, with additional medical therapy for persistent or recurrent symptoms.

• Initial treatment with invasive therapy, such as embolization or uterus-conserving surgery, with additional medical therapy for persistent or recurrent symptoms.

• Initial treatment with medical therapy, with additional invasive therapy (including hysterectomy) for persistent or recurrent symptoms.

• Initial treatment with invasive therapy, with additional invasive therapy (including hysterectomy) for persistent or recurrent symptoms.

"Recurrence" in the setting of conservative treatment for fibroids can mean several things:

• Regrowth of a fibroid after stopping suppressive therapy, such as GnRH agonists.

• Regrowth of a fibroid after incomplete surgical excision.

• Development of new fibroids.

For the purposes of the review, we considered all three possibilities, since most studies, especially of surgical treatments, did not distinguish between these possibilities.

During our review, we discovered that many studies reported "recurrence" rates after uterus-conserving treatment but did not often distinguish between anatomical recurrence documented by physical or radiological examination and symptomatic recurrence. Therefore, we summarize all recurrences below and, where possible, attempt to distinguish between anatomical and symptomatic recurrence. If possible, we also report on any patient or clinical characteristics that predicted either recurrence or the need for additional therapy.

Results

We did not identify any studies that allowed estimation of the incidence of additional medical treatment after either initial medical treatment or initial invasive therapy. Results for studies that reported either recurrence or the need for additional invasive therapy are summarized below.

No Treatment, Hormonal Therapy, or Treatment with NSAIDs

Incidence of recurrence

Table 15. No treatment or noninvasive therapy (other (more...)

Table 15. No treatment or noninvasive therapy (other than GnRH agonists)

Study	No. of patients	Length of followup	Diagnostic recurrence (bimanual exam or ultrasound)	Symptomatic recurrence/ persistence	Additional surgery other than hysterectomy	Hysterectomy
Carlson, Miller, and Fowler, 1994	106 (68% no treatment, 18% NSAIDs, 14% hormones)	12 months	Not reported	Bleeding: 83% Pain: 52% Fatigue: 73% New problems: 10%	Not reported	23% (overall; data not reported separately for patients with fibroids)

The Maine Women's Health Study (Carlson, Miller, and Fowler, 1994b) is the only prospective study we identified that used validated outcome measures and reported on the incidence of additional treatment in women followed expectantly or managed with nonsteroidal or hormonal medication (Table 15). Data on the 1-year incidence of hysterectomy were not reported separately for patients with fibroids, but the authors stated that this risk was not different among women with fibroids than among women with abnormal bleeding or pelvic pain. Data were not reported on additional treatments other than hysterectomy (e.g., the addition of medications to women initially managed expectantly or myomectomy).

Predictors of recurrence

Baseline data on uterine size, number of fibroids, and other anatomical measures were not reported. The association of other potential predictors of hysterectomy -- such as age, race, and parity -- with the occurrence of hysterectomy also was not reported.

GnRH Agonist Therapy

Incidence of recurrence

Table 16. Recurrence after GnRH agonist therapy

Table 16. Recurrence after GnRH agonist therapy

Study	No. of patients	Length of followup	Diagnostic recurrence (bimanual exam or ultrasound)	Symptomatic recurrence/ persistence	Additional surgery other than hysterectomy	Hysterectomy
Serra, Panetta, Colosimo, et al., 1992	110	12 months	Not reported	24.3% (n = 82, 7% lost to follow-up)	4 months: 8.4% 12 months: 6.8%	Not reported
van Leusden, 1992	22 pre-menopaus-al, 6 post-meno-pausal	Up to 42 months	Not reported	Not reported	Premenopausal: 4% Postmenopausal: 0%	Pre-menopausal: 55% Post-menopausal: 0%
Vollenhoven, Shekleton, McDonald, et al., 1990	40	Mean 9.6 months	Not reported	Not reported	22.5%	0
West, Lumsden, Hillier, et al., 1992	Buserelin + MPA, either combined (n = 10) or sequential (n = 10)	24 months	Not reported	Combined: 11% Sequential: 10%	Not reported	Combined: 33% Sequential: 30%

Although several studies reported on fibroid regrowth after cessation of GnRH agonist therapy (de Aloysio, Altieri, Pretolani, et al., 1995; Donnez, Schrurs, Gillerot, et al., 1989; Palomba, Affinito, Di Carlo, et al., 1999; Serra, Panetta, Colosimo, et al., 1992; van Leusden, 1992), fewer reported on symptomatic recurrence or incidence of surgical treatment (Table 16). However, given both the small number of studies that report data on symptoms and additional treatment and the small size of the studies, accurate estimation of the risk of symptomatic recurrence is difficult. Given that regrowth is an almost universal occurrence after cessation of GnRH agonist therapy in premenopausal women, it is not surprising that recurrent symptoms, in some cases leading to myomectomy or hysterectomy, occur as well. However, it is unclear in what proportion of women symptoms recur, at what level of severity, and whether alternative treatments would be more or less likely to work after treatment with a GnRH agonist.

Predictors of recurrence

In the small series described by van Leusden (van Leusden, 1992), perimenopausal women appeared less likely to undergo surgery after treatment; however, the small size of the study precluded significance testing. None of the other studies reported on factors influencing the likelihood of recurrent symptoms or incidence of subsequent surgery.

Myomectomy

Incidence of recurrence

Table 17. Recurrence after myomectomy

Table 17. Recurrence after myomectomy

Study	No. of patients	Length of followup	Diagnostic recurrence/ persistence (bimanual or ultrasound)	Symptomatic recurrence/ persistence	Additional surgery other than hysterectomy	Hysterectomy
Abdominal myomectomy
Berkeley, De Cherney, and Polan, 1983	50	Median 50.7 months	Not stated	Not stated	8% laparoscopy for infertility	8%
Egwuatu, 1989	55	1-4 years	Not reported	Not reported	10.5%	3.1%
Fedele, Parazzini, Luchini, et al., 1995	145	60 months	5-year cumulative probability: 51%	Not reported	Not reported	Not reported
Gehlbach, Sousa, Carpenter, et al., 1993	37 (infertility patients)	Minimum of 12 months	47%	Not reported	10.8%	5.4%
Mais, Ajossa, Guerriero, et al., 1996	20 abdominal, 20 laparo-scopic	6 months	Abdominal: 5%, Laparoscopic: 10%	Not reported	Not reported	Not reported
Sirjusingh, Bassaw, and Roopnarinesingh, 1994	114	2-10 years	Not reported	17%	Myomectomy: 4%	Hysterectomy: 8%
Stringer, Walker, and Meyer, 1997	49 abdominal, 49 laparo-scopic	Not reported	Not reported	Not reported	Not reported	Abdominal: 6% Laparoscopic: 2%
Abdominal myomectomy with adjunctive GnRH
Fedele, Vercellini, Bianchi, et al., 1990	GnRH: 8 No GnRH: 16		Not reported	GnRH: 62.5% No GnRH: 12.5% (OR 10.2; 95% CI, 1.6, 63.3)	Not reported	Not reported
Friedman, Daly, Juneau-Norcross, et al., 1992	GnRH: 9 No GnRH: 9	27-38 months	GnRH: 67% No GnRH: 56%	GnRH: 22.2% No GnRH: 22.2%	Not reported	Not reported
Sudik, Husch, Steller, et al., 1996	GnRH: 33 No GnRH: 34	12 months	46.3%	Not reported	Not reported	Not reported
Vavala, Lanzone, Monaco, et al., 1997	65: 40 no post-op treatment, 25 post-op GnRH (not random­ized)	36 months	No GnRH: 22.5% GnRH: 4%	Not reported	Not reported	Not reported
Hysteroscopic myomectomy
Bernard, Darai, Poncelet, et al., 2000	31	Minimum 12 months	Not reported	Not reported	6.5% repeat myomectomy	Not reported
Corson and Brooks, 1991	92	Mean 17 months	Not reported	Not reported	Overall: 17.3% Hysteroscopic myomectomy: 7.6% Abdominal myomectomy: 2.1% Endometrial ablation: 4.3%	3.2%
Cravello, D'Ercole, Boubli, et al., 1995	239	Mean 2.5 years	Not reported	Not reported	16% repeat myomectomy	Not reported
Cravello, Farnarier, De Montgolfier, et al., 1999	196	Mean 73 months	Not reported	35.7%	Myomectomy: 8.8% Ablation: 6.5%	14.6%
De Blok, Dijkman, and Hemrika, 1996	109	Mean 2.8 years	Not reported	Not reported	Repeat procedure: 11.9%	5.5%
Derman, Rehnstrom, and Neuwirth, 1991	108	1-16 years	Not reported	25%	Cumulative risk 37% after 11 years	6.5%
Dueholm, Forman, and Ingerslev, 1998	45 (38 with residual tissue at completion of procedure)	6 months	Not reported separately	20%	20% (23.6% with residual tissue)	2.2% (2.6% with residual tissue)
Emanuel, Wamsteker, Hart, et al., 1999	285	Median 46 months	Not reported separately	Not reported separately	76.7% at 8 years	10.8% at 8 years
Hallez, 1995	274	5-63 months	Not reported	33% after 60 months	Not reported	Not reported
Hart, Molnar, and Magos, 1999	122	Mean 2.3 years	Not reported	Not reported	21% first 4 years, none after 4 years	Not reported
Indman, 1993	38 with concurrent ablation, 13 without	Mean 2.2 years	Not reported	Not reported	Not reported	With ablation: 0 Without ablation: 8%
Itzkowic, 1993	40	Mean 14 months	Not reported	Not reported	7.5%	2.5% converted at time of initial surgery
Lin, Iwata, and Liu, 1994	25 (pretreated with GnRH)	9 months	Not reported	12%	13.6%	Not reported
Loffer, 1990	43	Minimum 12 months	Not reported	Not reported	Repeat hysteroscopy: 9.4% Other: 5.6%	9.4%
Mints, Radestad, and Rylander, 1998	62	Mean 29 months	Not reported	Not reported	Not reported	12%
Phillips, Nathanson, Meltzer, et al., 1995	208 (120 with myo-mectomy alone, 88 with myo-mectomy + endometrial ablation)	6 months (n = 208); 6 years (n = 185)	Not reported	Myomectomy: 15% Myomectomy + ablation: 11.5%	Myomectomy: 4.3% Myomectomy + ablation: 3.2%	Myomectomy: 0.5% Myomectomy + ablation: 1.6%
Vercellini, Zaina, Yaylayan, et al., 1999	108	Mean 41 months	34% (3 years)	30% (3 years)	Not reported	Not reported
Laparoscopic myomectomy
Chapman, 1998	300	6-72 months	Not stated	Not stated	Additional laser treatment: 10%		2%
Nezhat, Roemisch, Nezhat, et al., 1998	114	Mean 37 months	33.3% Cumulative risk: 1 year 10.6% 3 year 31.7% 5 year 51.4%	Not reported	12.3%		6.1%

The reported incidence of recurrence of fibroids -- as defined by anatomical diagnosis, recurrence of symptoms, need for any additional surgery, or hysterectomy after abdominal, hysteroscopic, and laparoscopic myomectomy -- is shown in Table 17.

Given variation in definitions of recurrence and length of followup, estimation of an overall probability of symptomatic recurrence after myomectomy is not possible. Reported incidence of subsequent conservative surgery ranged from 10-12 percent for laparoscopic myomectomy, 4-10 percent for abdominal myomectomy, and 3-76 percent for hysteroscopic myomectomy. Incidence of hysterectomy ranged from 2-6 percent for laparoscopic myomectomy, 3-8 percent for abdominal myomectomy, and 0.5-12 percent for hysteroscopic myomectomy. Given the likely differences in patient characteristics that led to the selection of a particular treatment in these nonrandomized studies, comparison of rates between different approaches to myomectomy is inappropriate.

Predictors of recurrence

Patient characteristics associated with a statistically significant increased risk of recurrence in individual studies included:

• Increasing preoperative uterine size (Emanuel, Wamsteker, Hart, et al., 1999) (hysteroscopic myomectomy only).

• Increasing number of fibroids (Emanuel, Wamsteker, Hart, et al., 1999; Fedele, Parazzini, Luchini, et al., 1995; Friedman, Daly, Juneau-Norcross, et al., 1992; Nezhat, Roemisch, Nezhat, et al., 1998).

• Incomplete excision of fibroid (Emanuel, Wamsteker, Hart, et al., 1999) (hysteroscopic myomectomy only) (Dueholm, Forman, and Ingerslev, 1998).

• Increasing extension into the myometrium of submucous fibroids (Nezhat, Roemisch, Nezhat, et al., 1998) (hysteroscopic myomectomy only) (Emanuel, Wamsteker, Hart, et al., 1999).

Embolization

Incidence of recurrence

Table 18. Recurrence after uterine artery embolization (more...)

Table 18. Recurrence after uterine artery embolization

Study	No. of Patients	Length of followup	Ultrasonographic recurrence/ persistence	Symptomatic recurrence/ persistence	Additional surgery/ procedure	Hysterectomy
Goodwin, McLucas, Lee et al., 1999	60	Mean 16.3 months	12%	12%	13%	10%
Hutchins, Worthington-Kirsch, and Berkowitz, 1999	305	Up to 12 months	Not reported	3 months: 13%, 6 months: 13% 12 months: 14% (crude rate; variable loss to followup, followup time prohibits calculation of cumulative rate)	4.2% (crude rate; variable loss to followup, followup time prohibits calculation of cumulative rate)	2.0% (crude rate; variable loss to followup, followup time prohibits calculation of cumulative rate)
Pelage, Le Dref, Soyer, et al., 2000	80	2 years	Not reported	6% "no improvement"	6%	1.2%
Spies, Scialli, Jha, et al., 1999	61	Mean 8.7 months	Not reported	5% ("dissatisfied")	5%	3.4%

Reported experience with uterine artery embolization is limited in terms of both numbers of patients followed and length of followup (Table 18). In addition, in the paper with the largest number of patients (Hutchins, Worthington-Kirsch, and Berkowitz, 1999), loss to followup was substantial, making estimations of true cumulative rates of recurrence, persistence, or need for subsequent therapy difficult.

Predictors of recurrence

None were reported.

Methodological Issues

The available literature does not allow calculation of summary estimates of the need for additional treatment after uterus-sparing interventions. Limitations of the literature that preclude summary estimates include:

• Variability in surgical technique: Even for a given procedure, such as abdominal myomectomy, it is possible that variation in technique might affect recurrence risk.

• Variability in definitions of recurrence: Some studies reported radiological persistence or recurrence without correlation with symptoms; others reported only those patients needing additional surgical treatment. We did not identify any study that addressed the issue of the use of medical treatments such as hormonal or nonsteroidal agents after myomectomy or embolization.

• Variability in methodology used to estimate recurrence risk: Some studies reported only the number of patients requiring additional treatment without accounting for loss to followup or time of additional treatment, while others appropriately used life-table measures.

• Variability in reporting of potential predictors of risk: Considerable variation was observed in the reporting of the size, location, and number of fibroids removed, as well as other variables such as age, race, parity, and educational level.

• Variability in patient populations: Women undergoing uterus-sparing therapy who plan on pregnancy represent a different population than women who have no plans for future pregnancies; the type of additional therapies available are quite different, so "recurrence" as defined by additional treatment may not reflect symptomatic recurrence or persistence.

• Variable length of followup: Variable length of followup makes estimation of the true risk of recurrence or persistence difficult, unless survival analysis techniques are explicitly used to account for this variability.

• Variable loss to followup: Patients lost to followup may be more likely to have recurrent symptoms and seek treatment elsewhere. Although life-table methods may be helpful in accounting for the effects of this loss on estimates of cumulative recurrence, these methods depend on an assumption of nondifferential treatment outcomes between patients who remain in the study and those who are lost to followup.

• "Recurrence" versus new fibroids: Given the relatively high incidence of fibroids, it is possible that at least some fibroids detected after uterus-conserving treatment represent new lesions that would have developed with or without the intervention. Indeed, it is possible that removing fibroids may induce changes, such as increased expression or secretion of cytokines and growth factors that facilitate the growth of new fibroids. Thus, development of fibroids after conservative therapy may not represent failure of the therapy but may simply reflect the natural history of uterine leiomyomata.

Summary

Clearly, all uterus-sparing treatments for symptomatic fibroids leave some risk of recurrence of symptoms, resulting in the need for additional therapy. However, the available literature does not allow quantification of this risk for any of these "conservative" treatments. For myomectomy, factors such as size, number, and location of fibroids appear to increase the risk of recurrence. It is unclear whether this increased risk is related to differences in underlying biology (women with larger, more numerous fibroids may be more likely to develop new fibroids) or the technical difficulty of the surgery. Future estimation of the risk of recurrence and need for additional treatment would be facilitated by standardization in the reporting of pre- and postoperative patient characteristics, definitions of recurrence and persistence, and use of appropriate statistical techniques. Without such standardization, quantitative comparison of recurrence risk between treatments is impossible.

Approach

As with Question 4, we approached this question by dividing it into four separate clinical scenarios describing the initial treatment. We then considered three scenarios in which increased morbidity might result from additional treatment:

• Initial treatment with medical therapy, with additional medical therapy for persistent or recurrent symptoms; or initial treatment with invasive therapy such as embolization or uterus-conserving surgery, with additional medical therapy for persistent or recurrent symptoms. Theoretically, increased morbidity could result in such circumstances from either side effects specifically related to the additional agents, drug interactions when additional pharmaceutical agents are added to a preexisting drug regimen, or changes in physiology related to prior invasive therapy, such as embolization or surgery, which would lead to either increasing the severity of side effects or new types of side effects.

• Initial treatment with medical therapy, with subsequent invasive management of recurrent or persistent symptoms. Theoretically, morbidity could result from steroid hormones. Estrogens appear to increase the risk of perioperative venous thrombolic events (Anonymous. 1999), and current practice is to discontinue oral contraceptives and other estrogens for at least 4 weeks prior to surgery. It is conceivable that patients using oral contraceptives to treat symptoms related to fibroids who did not discontinue these agents prior to additional surgical therapy might have an increased risk of venous thrombophlebitis or pulmonary embolism. Nonsteroidal anti-inflammatory drugs (NSAIDs) may increase the risk of perioperative hemorrhage by inhibiting platelet function (Connelly and Panush, 1991; Scher, 1996); again, current practice is to discontinue their use prior to procedures with substantial bleeding risk. Patients using NSAIDs to treat symptoms related to fibroids who did not discontinue these agents prior to additional invasive therapy might have an increased risk of hemorrhagic complications, including hematomas. In the case of changes in underlying severity of illness or development of comorbidity, it is possible that medical therapy that alleviates symptoms without substantially altering the natural history of fibroid growth could result in increased risks of morbidity with subsequent therapy, either by allowing further growth of the fibroids -- resulting in an increased risk of surgical morbidity (Hillis, Marchbanks, and Peterson, 1996) -- or by the development of unrelated conditions (e.g., diabetes, hypertension, or cardiac disease) that would increase the risk of complications.

• Initial treatment with invasive therapy, with additional invasive therapy (including hysterectomy) for persistent or recurrent symptoms might increase the risk of morbidity in several ways. One is through multiple exposure to risk. Even if having undergone a prior procedure does not increase the inherent risk of morbidity associated with the procedure, patients undergoing additional procedures will have an increased likelihood of a complication or adverse outcome simply by being exposed to the same risks on multiple occasions. These multiple exposures increase the cumulative risk of an adverse outcome (although for most serious complications, this cumulative risk is likely to be quantitatively small). Changes in the underlying severity of condition or development of comorbidity also might increase the risk of morbidity. As with medical therapy, it is possible that the interval development of unrelated conditions might increase the risk of subsequent invasive procedures. Increased technical difficulty related to a prior procedure also may increase risk. Any surgical procedure can result in adhesions, and the association of myomectomy with adhesion formation is well-documented (Adamson, 1993; Anonymous. 1992; Diamond, Bieber, Coddington, et al., 1996; Franklin, Haney, Kettel, et al., 1995; Haney, 1997; Mais, Ajossa, Piras, et al., 1995; March, Boyers, Franklin, et al., 1993; Murray and Tulandi, 1996; Tulandi, Murray, and Guralnick, 1993; Ugur, Turan, Mungan, et al., 1996). The presence of adhesions increases the technical difficulty of the procedure and may increase the risk of complications.

Results

Methodological Issues

Even if additional treatment does result in additional risk (especially for surgical procedures), this additional risk needs to be considered in the context of the likelihood of successful initial treatment when counseling individual patients. For example, consider an initial noninvasive treatment with a likelihood of success of 75 percent with 25 percent of patients requiring a second procedure and a complication rate of 5 percent. The overall risk of a complication for all women undergoing the initial treatment is 5 percent + (25 percent x 5 percent), or 6.25 percent. Therefore, the effect of an increased cumulative risk must be considered in the context of the likelihood of additional therapy.

Summary

There are no data to allow estimation of the degree to which additional therapy, especially additional invasive therapy, increases the overall risk of morbidity for an individual patient or whether any increased risk is statistically or clinically significant.

Approach

In this section we consider nonsurgical approaches to managing fibroids, including observation (no treatment), estrogens, progestins, NSAIDs, GnRH agonists, and other drug treatments. These drugs are used either as primary treatments or (in the case of GnRH agonists) as adjuncts to surgical treatment. We did not evaluate possible primary prevention strategies; these are discussed in Chapter 6 (Future Research).

Results

We identified 70 separate studies of medical treatments for uterine fibroids (see Evidence Tables 2 and 3). These included nine studies describing the effects of no treatment (natural history), two studies of NSAIDs, three studies of estrogen-progestin combinations, six studies of progestins, 52 studies of GnRH agonists, and six studies of other hormonal and miscellaneous treatments. All study populations consisted of pre- or perimenopausal women. Nineteen of the GnRH agonist studies were uncontrolled case series; the remainder had control groups, and most of these allocated patients randomly to treatment and control groups. Four of the 20 studies of other drugs were uncontrolled case series.

No Treatment

The results of conservative management (no treatment) were described in the placebo arms of four randomized controlled trials of hormone treatment that followed patients for long periods (more than 6 months) (Friedman, Harrison-Atlas, Barbieri, et al., 1989; Friedman, Hoffman, Comite, et al., 1991; Gregoriou, Vitoratos, Papadias, et al., 1997; Schlaff, Zerhouni, Huth, et al., 1989) and in one cohort study in which the majority of nonsurgically managed patients received observation alone (Carlson, Miller, and Fowler, 1994b).

One of the most useful studies comparing nonsurgical with surgical management of fibroids was a large cohort study (Carlson, Miller, and Fowler, 1994b). A cohort of women with fibroids (uterine size 8 weeks or greater) was assembled from multiple sites in Maine and interviewed prospectively. One hundred and twenty-three women were managed with hysterectomy, and 106 received one of a variety of nonsurgical treatments (observation alone, 68 percent; NSAIDs or iron, 18 percent; hormone therapy, 14 percent). There were significant differences between the two groups at baseline. The patients who were managed without surgery had fewer days of bleeding and dysmenorrhea than those who underwent hysterectomy. In addition, the patients managed without surgery had more positive feelings about their symptoms and better quality of life at baseline than those who underwent surgery. By 1 year, the women who underwent hysterectomy had increased quality-of-life scores and, compared with nonsurgical patients, felt better about their symptoms. Approximately 23 percent of nonsurgical patients underwent hysterectomy in the 1 year followup period.

Baseline differences in the symptoms, attitudes, and quality of life between women managed nonsurgically and those managed surgically suggest that women who were managed nonsurgically were less severely affected by their fibroids. Because of this imbalance, we are limited in our ability to draw inferences about the relative effectiveness of nonsurgical and surgical management from this study. Furthermore, because nonsurgical management included several different treatment approaches and results were not reported separately for the various interventions, comparisons among the various nonsurgical treatments used are impossible.

Table 19. Natural history of uterine fibroids in prospective controlled (more...)

Table 19. Natural history of uterine fibroids in prospective controlled studies on no treatment or placebo

Study	No. of patients	Length of followup	Baseline uterine or fibroid size	Followup uterine or fibroid size	Change
Friedman, Harrison-Atlas, Barbieri, et al., 1989	Placebo (n = 20)	6 months	Uterine size: 426 ± 43 cc	Uterine size: 429 ± 52 cc	No change (p = not significant)
Friedman, Hoffman, Comite, et al., 1991	Placebo (n = 65)	6 months	Uterine size: 452 ± 51 cc	Uterine size: 5% increase	No change (p = not significant)
Gregoriou, Vitoratos, Papadias, et al., 1997	No treatment (n = 20)	12 months	Fibroid size: 118.4 cc	Fibroid size: 117.5 cc	No change in fibroid volume
Schlaff, Zerhouni, Huth, et al., 1989	Placebo (n = 6)	6 months	Uterine size: 457 ± 106 cc Fibroid size: 267 ± 82 cc	Uterine size: 656 ± 208 Fibroid size: 417 ± 169 cc	No change (p = not significant)

Long-term observation (at least 1 year) or no treatment was described in the control arm of a trial of the synthetic steroid tibolone, a drug with mixed estrogenic, progestogenic, and androgenic properties (Gregoriou, Vitoratos, Papadias, et al., 1997). In addition, several placebo-controlled studies of the GnRH agonist leuprolide described 6-month outcomes of placebo-treated patients (Friedman, Harrison-Atlas, Barbieri, et al., 1989; Friedman, Hoffman, Comite, et al., 1991; Schlaff, Zerhouni, Huth, et al., 1989) (see Table 19). All of these studies were conducted among women who did not want to have surgery (Friedman, Harrison-Atlas, Barbieri, et al., 1989; Friedman, Hoffman, Comite, et al., 1991; Gregoriou, Vitoratos, Papadias, et al., 1997; Schlaff, Zerhouni, Huth, et al., 1989). Two described hemoglobin or hematocrit decreases over time, in addition to changes in fibroid or uterine size (Friedman, Harrison-Atlas, Barbieri, et al., 1989; Friedman, Hoffman, Comite, et al., 1991). None of the studies reported the number of women who eventually underwent surgery.

Consistent with the Maine Women's Health Study, in which the majority of nonsurgically treated patients received no treatment, these studies showed that "no treatment" resulted in no reduction in uterine size, no alleviation of symptoms, and no resolution of anemia.

Hormone Replacement Therapy with Estrogen and Progestins

Hormone replacement therapy (HRT) with estrogen-progestin was compared with the synthetic steroid tibolone in one study (de Aloysio, Altieri, Penacchioni, et al., 1998). HRT has also been studied as "add-back" therapy during or after treatment with GnRH agonists (Friedman, Daly, Juneau-Norcross, et al., 1994; Maheux and Lemay, 1992).

Tibolone (2.5 mg) was compared with a combination of conjugated equine estrogen (0.625 mg) and medroxyprogesterone acetate (MPA) (5 mg) daily for 1 year (de Aloysio, Altieri, Penacchioni, et al., 1998). Neither group of 25 patients showed a significant change in fibroid size by transvaginal ultrasound, but the proportion of cycles with vaginal bleeding or spotting was significantly lower in the tibolone-treated patients than in the HRT-treated patients (p < 0.02).

Several studies have tested the efficacy of low-dose hormone therapy added to GnRH agonist treatment, referred to as an "add-back" regimen by some authors (Friedman, Daly, Juneau-Norcross, et al., 1993). One study compared an estrogen-progestin with a progestin-only add-back regimen among 51 patients who also were treated with leuprolide acetate for 2 years (Friedman, Daly, Juneau-Norcross, et al., 1994). The add-back regimen was initiated 3 months after the GnRH agonist was started and continued for 21 months. In this study, uterine volume decreased during the 3-month period of GnRH agonist-only therapy by 36-44 percent. Once estrogen-progestin or progestin-only therapy was instituted, uterine volume remained the same in the combined estrogen-progestin group but increased significantly in the progestin-only group. The two groups had similar increases in hemoglobin and hematocrit. The effect on serum high-density lipoprotein (HDL) cholesterol concentration favored the estrogen-progestin group, but there were no between-group differences in bone density of the lumbar spine. The authors of this study concluded that prolonged (more than 6 months) medical treatment with estrogen-progestin combination is safe and effective.

Another study compared sequential versus continuous cyclical low-dose hormone replacement therapy (conjugated equine estrogen and MPA) given as "add-back" therapy along with goserelin depot (Maheux and Lemay, 1992). Mean fibroid volume decreased significantly during the first 3 months of GnRH agonist treatment, but no further changes were evident after hormone replacement therapy was instituted until after GnRH agonist therapy was stopped. Significantly less bleeding was observed in the group receiving continuous rather than sequential hormone replacement therapy. Similarly, hot flashes were less severe in patients receiving continuous HRT than in those receiving sequential HRT.

Our search identified one randomized study of oral contraceptives (Friedman and Thomas, 1995), which was subsequently withdrawn without explanation.

Progestins

Table 20. Effect of medroxyprogesterone acetate therapy (more...)

Table 20. Effect of medroxyprogesterone acetate therapy in conjunction with GnRH agonist treatment of uterine fibroids

Study	Hormonal therapy	GnRH co-intervention (timing)	Baseline uterine volume	Post-treatment uterine volume	Results
Benagiano, Morini, Aleandri, et al., 1990	MPA 200 mg tapering to 25 mg daily (n = ?) No treatment (n = ?)	Buserelin (concurrent)	132 cc 122 cc	170 cc 198 cc	No difference
Caird, West, Lumsden, et al., 1997	MPA 15 mg qd (n = 12) Placebo (n = 12)	Goserelin (concurrent)	-	-	No difference between groups
Carr, Marshburn, Weatherall, et al., 1993	MPA 20 mg qd 1st 3 mo of GnRH Rx (n = 8) MPA 20 mg qd 2nd 3 mo of GnRH Rx (n = 8)	Leuprolide (concurrent)	100% 100%	78% 74%	No difference between groups
Friedman, Barbieri, Doubilet, et al., 1988	MPA 20 mg qd (n = 9) Placebo (n = 7)	Leuprolide (concurrent)	811 cc 601 cc	688 cc 294 cc*
Scialli and Jestila, 1995	MPA 5 mg qd (n = 21) Placebo (n = 20)	Leuprolide (before)	538 cc 344 cc	473 cc 510 cc
West, Lumsden, Hillier, et al., 1992	MPA 15 mg qd (n = 10) MPA 15 mg qd after 3 months (n = 10)	Goserelin (concurrent)	100% 100%	82% 77%	No difference

* P< 0.05 difference from baseline to posttreatment

Five of the studies of MPA given either concurrently with or following GnRH agonist treatment used doses of MPA ranging from 5 mg to 20 mg daily; however, one study involved treatment with 200 mg daily, tapering to 25 mg daily over 4 months (Benagiano, Morini, Aleandri, et al., 1990). In none of these studies did progestin add-back therapy have an effect on uterine size. Symptoms were reported in markedly different ways among these trials. Vasomotor symptoms, such as hot flashes, were decreased with MPA treatment in both placebo-controlled studies that reported this outcome (Caird, West, Lumsden, et al., 1997; Friedman, Barbieri, Doubilet, et al., 1988), while MPA treatment was associated with increased spotting (irregular bleeding) in two studies (Benagiano, Morini, Aleandri, et al., 1990; Friedman, Barbieri, Doubilet, et al., 1988) and no difference in two studies (Caird, West, Lumsden, et al., 1997; Scialli and Jestila, 1995).

Nonsteroidal Anti-inflammatory Drugs

Two studies described the effects of NSAIDs taken during menstruation to reduce menorrhagia associated with fibroids. The agents studied were indomethacin (Anteby, Yarkoni, and Ever Hadani, 1985) and naproxen (Ylikorkala and Pekonen, 1986). Both studies used a randomized design and included a placebo control. Both included women with menorrhagia from a variety of causes but reported results for the fibroid subgroup separately. One study found statistically significant reductions in days of bleeding (Anteby, Yarkoni, and Ever Hadani, 1985), but there were no differences in mean blood loss during menses in the other study (Ylikorkala and Pekonen, 1986).

GnRH Agonists as Primary Treatment

Fifty-four studies examined the use of GnRH agonists alone. In addition, nine studies examined the effectiveness of combining GnRH agonist treatment with progesterone, estrogen-progestin combinations, or danazol.

Most of these studies described the short-term reduction in uterine or fibroid volume observed with GnRH agonist treatment. The average percentage reduction in uterine size in the 21 studies reporting percentage reduction or pre- and posttreatment values was 46.7 percent (standard deviation [SD] 12.4 percent). Fewer studies measured fibroid volume. Studies that reported both total uterine volume and fibroid volume found the percentage of reductions to be similar (Broekmans, Hompes, Heitbrink, et al., 1996; Costantini, Anserini, Valenzano, et al., 1990; Palomba, Affinito, Di Carlo, et al., 1999), suggesting that shrinkage occurs in fibroid tumors, as well as in the nonmyomatous uterus.

Most of the reduction in fibroid and uterine volume occurred during the first month of treatment with the GnRH agonist, with diminishing reductions in size over subsequent months. The magnitude of change in uterine size during the first month was found to be a significant predictor of the ultimate response (Hackenberg, Gesenhues, Deichert, et al., 1992).

Several studies continued to monitor uterine or fibroid size after GnRH agonist therapy was discontinued (de Aloysio, Altieri, Pretolani, et al., 1995; Donnez, Schrurs, Gillerot, et al., 1989; Palomba, Affinito, Di Carlo, et al., 1999; Serra, Panetta, Colosimo, et al., 1992; van Leusden, 1992). These studies found that after GnRH agonists were stopped, uterine and fibroid sizes returned toward pretreatment values over several months.

Table 21. Symptomatic response to GnRH agonist treatment (more...)

Table 21. Symptomatic response to GnRH agonist treatment of uterine fibroids

Study	Year	Initial symptoms	Proportion of patients with resolution of symptoms
Coddington, Brzyski, Hansen, et al., 1992	1992	Menorrhagia (70%), urinary frequency (50%), and tight clothes (50%)	100%
Friedman, Barbieri, Benacerraf, et al., 1987	1987	Pelvic pain, mass, or menorrhagia	100%
Friedman, Hoffman, Comite, et al., 1991	1991	Bloating, menorrhagia, pelvic pressure, pelvic pain, constipation, urinary frequency, dyspareunia, menometrorrhagia	Not reported Symptom response was "better" in GnRH than placebo (no test of significance performed)
Nakamura, Yoshimura, Yamada, et al., 1991	1991	Bleeding	Amenorrhea: 24% Improved: 68% Severe bleeding episode: 8%
		Dysmenorrhea	100% resolved
		Pelvic pressure	100% resolved
Palomba, Affinito, Di Carlo, et al., 1999	1999	Not reported	Not reported There was a statistically significant reduction in intensity for all myoma related symptoms at sixth month of treatment (p < 0.01); this reduction persisted during the entire course of treatment.
Palomba, Affinito, Tommaselli, et al., 1998	1998	Menorrhagia, pelvic pressure, pelvic pain, constipation, urinary frequency	100% 94% 89% 100% 84%
Schlaff, Zerhouni, Huth, et al., 1989	1989	Pain	10/11 resolved, 7/11 persistent response 6 months after cessation
		Bleeding	7/7 resolved, 3/7 persistent response
		Pelvic pressure	6/8, 3/8 persistent response
		Constipation	4/6, 3/6 persistent response
		Dyspareunia	4/5, 2/5 persistent response
Serra, Panetta, Colosimo, et al., 1992	1992	Not reported	End of 4 months (n = 100): Resolved 34% Improved 60% No change 5.6%
Vollenhoven, Shekleton, McDonald, et al., 1990	1990	Not reported	"All" patients reported subjective improvement in symptoms.

Six studies reported data on the number of women with common symptoms but did not relate these to pretreatment symptoms (Abramovici, Dirnfeld, Auslander, et al., 1994; Cirkel, Ochs, Schneider, et al., 1992; Costantini, Anserini, Valenzano, et al., 1990; de Aloysio, Altieri, Pretolani, et al., 1995; Fedele, Bianchi, Raffaelli, et al., 2000; Watanabe, Nakamura, Matsuguchi, et al., 1992). Thirteen studies reported no data on fibroid-related symptom outcomes (Bianchi, Costantini, Anserini, et al., 1989; Broekmans, Hompes, Heitbrink, et al., 1996; Cagnacci, Paoletti, Soldani, et al., 1994; D'Addato, Repinto, and Andreoli, 1992; De Leo, Morgante, Lanzetta, et al., 1997; Donnez, Schrurs, Gillerot, et al., 1989; Felberbaum, Germer, Ludwig, et al., 1998; Friedman, Harrison-Atlas, Barbieri, et al., 1989; Golan, Bukovsky, Schneider et al., 1989; Hackenberg, Gesenhues, Deichert et al., 1992; Kuhlmann, Gartner, Schindler, et al., 1997; Ueki, Okamoto, Tsurunaga, et al., 1995; van Leusden, 1992). Furthermore, the recording and reporting of symptoms and response of symptoms to treatment were inconsistent.

Two studies assessed recurrence of symptoms after GnRH agonist treatment was discontinued (Palomba, Affinito, Di Carlo, et al., 1999; Serra, Panetta, Colosimo, et al., 1992). One study reported only that there was a significant increase in the intensity of myoma-related symptoms during treatment (Palomba, Affinito, Di Carlo, et al., 1999), but the other quantified the proportion of women with recurrent symptoms (Serra, Panetta, Colosimo, et al., 1992). Although all women were symptomatic before treatment, 94 percent had complete resolution or improvement of their symptoms after 4 months of treatment with leuprolide. At 8-12 months after treatment was discontinued, 24 percent of women had recurrent symptoms, 64 percent remained asymptomatic, and 7.3 percent were lost to followup.

Preoperative Use of GnRH Agonists

The most thoroughly studied use of medical treatments for fibroids is short-term treatment with GnRH agonists in preparation for hysterectomy or myomectomy (see Evidence Table 3). Nineteen randomized controlled trials on this topic were described in a recent Cochrane review and meta-analysis (Lethaby, Vollenhoven, and Sowter, 1999). In addition, we identified two randomized controlled trials (Cetin, Vardar, Demir, et al., 1995; Ylikorkala, Tiitinen, Hulkko, et al., 1995) not described in the Cochrane review, one of which was known to the review's authors but not yet assessed (Ylikorkala, Tiitinen, Hulkko, et al., 1995), and three nonrandomized studies with historical or concurrent controls describing the short-term use of GnRH agonists in preparation for definitive surgery (hysterectomy or myomectomy) (Falsetti, Mazzani, Rubessa, et al., 1992; Kiltz, Rutgers, Phillips, et al., 1994; Vercellini, Bocciolone, Colombo, et al., 1993).

Briefly, the Cochrane review found, consistent with the nonsurgical studies, that GnRH analog therapy given for 3-4 months prior to surgery significantly reduced uterine volume and uterine size and improved preoperative hemoglobin and hematocrit. Pelvic symptoms also were reduced, but some adverse events were more likely during GnRH agonist therapy.

Operating time and duration of hospital stay were reduced in GnRH agonist-treated patients. Also, more of the women undergoing hysterectomy were able to have a vaginal rather than an abdominal approach. The intraoperative estimated blood loss (EBL) and rate of vertical incisions were reduced for both myomectomy and hysterectomy. There were no significant differences in transfusion rates. There were no data with which to assess the effects of pretreatment with GnRH agonists on postmyomectomy fertility.

Two randomized controlled trials were not included in the Cochrane review. Ylikorkala, Tiitinen, Hulkko, et al. (1995) described a large trial comparing nafarelin delivered intranasally with a placebo nasal spray. During the 3-month pretreatment period, the nafarelin group experienced a 23.7 percent decrease in uterine size, a 31.4 percent decrease in fibroid size, and a 5.5 mg/dl increase in hemoglobin. However, the study found no significant difference in intraoperative outcomes of estimated blood loss or operating room time. Although the operating room time results are consistent with those reported in the Cochrane review, the EBL data are not. Notably, the change in uterine size in this study was somewhat smaller than that described in other studies and may explain the negative findings.

Cetin, Vardar, Demir, et al. (1995) describe a smaller trial involving 30 women with symptomatic fibroids who were randomized to buserelin intranasally for 3 months prior to myomectomy or immediate myomectomy. The investigators observed a 53 percent decrease in uterine size (p < 0.05) and an increase in hemoglobin from 10.5 to 13.4 (p < 0.05) from pre- to postbuserelin treatment. In this study, EBL was significantly lower in the GnRH agonist group than in the control group (135 cubic centimeters (cc) versus 292 cc; p < 0.05). Operating room time also was significantly lower in the GnRH agonist group than in the control group (87 minutes versus 102 minutes; p < 0.05).

Table 22. Nonrandomized studies of preoperative GnRH (more...)

Table 22. Nonrandomized studies of preoperative GnRH agonists on operative outcomes

Study	Presurgical treatment	Type of surgery	Uterine volume decrease in GnRH-treated patients	Intraoperative blood loss (ml)	Operating room time (minutes)
Falsetti, Mazzani, Rubessa, et al., 1992	Goserelin 4 months (n = 30)	Myomectomy	48%	205 vs. 310 p < 0.001	Single: 91 vs. 97 Multiple: 111 vs. 118 p = not significant
Kiltz, Rutgers, Phillips, et al., 1994	Leuprolide for 3 months	Myomectomy	35%	680 vs. 722 p = not significant	Not reported
Vercellini, Bocciolone, Colombo, et al., 1993	Goserelin and iron for 6 months	Total abdominal hysterectomy (TAH) or vaginal hysterectomy (VH)	52%	186 vs. 351 p < 0.001	TAH: 102 vs. 104 VH: 83 vs. 77 p = not significant

Postoperative Use of GnRH Agonists

One study described the rate of recurrence of fibroids after myomectomy on no treatment or leuprolide (Vavala, Lanzone, Monaco, et al., 1997). Women who agreed to the treatment took leuprolide acetate (LA) depot 3.75 mg subcutaneous (SC) each month for 3 consecutive months each year for 3 years, starting 4 months after surgery. Women who refused postoperative leuprolide treatment were followed as a control group. Regrowth of fibroids (> 2 cm) and uterine size were both statistically significantly less likely in leuprolide-treated than control patients (p < 0.05). There was no assessment of symptom recurrence, fertility, or need for additional surgery, which limits the clinical relevance of the study.

Other Medical Treatments

Several studies described the use of other hormonally active drugs, including danazol (De Leo, la Marca, and Morgante, 1999; De Leo, Morgante, Lanzetta, et al., 1997; Ueki, Okamoto, Tsurunaga, et al., 1995), gestrinone (Coutinho, 1990; Coutinho, Boulanger, and Goncalves, 1986; Coutinho and Goncalves, 1989), tibolone (de Aloysio, Altieri, Penacchioni, et al., 1998; Gregoriou, Vitoratos, Papadias, et al., 1997), and RU-486 (Murphy, Morales, Kettel, et al., 1995). One study described the use of the Chinese herbal medicine kuei-chin-fu-ling-wan (Sakamoto, Yoshino, Shirahata, et al., 1992).

In an uncontrolled trial, De Leo, la Marca, and Morgante, et al. (1999) found that 4 months of treatment with danazol (400 mg per day) reduced uterine size and improved hematocrit. No data on the effect on fibroid symptoms were reported. In an earlier study, De Leo, Morgante, Lanzetta, et al. (1997) found that treatment with danazol (100 mg per day) after 6 months of GnRH agonist treatment reduced regrowth of fibroids by 31 percent compared with historical patients who received no danazol posttreatment (p < 0.001). Ueki, Okamoto, Tsurunaga, et al. (1995) compared change in fibroid size in patients taking danazol (400 mg daily) with patients taking the GnRH agonist buserelin. Although no statistical tests were reported, the study found a large difference in fibroid size reduction between danazol- and buserelin-treated patients (27 percent reduction versus 52 percent reduction, respectively).

Gestrinone, a synthetic steroidal hormone with androgenic, antiestrogenic, and antiprogestogenic properties, was studied in one placebo-controlled study (Coutinho, 1990). Among patients treated with gestrinone (5 mg three times per week), uterine size decreased after 2 months of treatment, and all patients reported alleviation of pretreatment symptoms attributed to fibroids. A high proportion of gestrinone-treated patients reported androgenic side effects, such as hirsutism (25 percent) and acne (87.5 percent). Two other studies tested the effectiveness of lower doses of gestrinone (2.5 mg three times per week) given by mouth or intravaginally (Coutinho, Boulanger, and Goncalves, 1986; Coutinho and Goncalves, 1989). The lower dose and intravaginal route of administration resulted in reductions in uterine size and fibroid-related symptoms similar to those achieved with the oral dose tested in the previously mentioned study.

The synthetic steroid tibolone, which has mixed estrogenic, progestogenic, and androgenic properties, was studied in two controlled trials. Gregoriou, Vitoratos, Papadias, et al. (1997) compared tibolone 2.5 mg daily versus no treatment in 40 women with at least one fibroid more than 20 mm in diameter. After 1 year of treatment, there was no significant change in fibroid volume in either group.

In the other study of tibolone, de Aloysio, Altieri, Penacchioni, et al. (1998) similarly found no change in fibroid size in patients treated for 1 year with 2.5 mg daily of tibolone or a separate group treated with conjugated equine estrogens 0.625 mg daily and medroxyprogesterone acetate 5 mg daily.

The antiprogestin agent RU-486 was tested in a prospective nonrandomized clinical trial comparing three doses (Murphy, Morales, Kettel, et al., 1995). Limited data showed that although all groups became amenorrheic, there was a smaller reduction in uterine size in the lowest dose group (5 mg daily) compared with the higher (25 mg or 50 mg daily) dose groups (25 percent reduction versus 50 percent and 55 percent reduction; no p-value reported).

A single study reported the results of treatment with the Chinese herbal medicine kuei-chin-fu-ling-wan (Sakamoto, Yoshino, Shirahata, et al., 1992). Kuei-chin-fu-ling-wan, known as keishi-bukuryo-gan (KBG) in Japan, contains five components: bark of Cinnamomum cassia Bl. (Lauraceae), root of Paeonia lactiflora Pall. (Paeoniaceae), seed of Prunus persica Batsch. or P. persiba Batsch.var.davidiana Maxim. (Rosaceae), carpophores of Poria cocos Wolf. (Polyporaceae), and root bark of Paeonia suffruticosa Andr. (Paeoniaceae). It has been reported to act as a luteinizing hormone-releasing hormone (LHRH) antagonist and a weak antiestrogen on uterine DNA synthesis in immature rats (Sakamoto, Kudo, Kawasaki, et al., 1988). Only women with uterine size 10 cm or smaller were included in the study. In addition to fibroid size, investigators assessed symptoms of menorrhagia (57 percent of patients) and dysmenorrhea (46 percent of patients). Fibroids showed complete resolution in 19 percent of patients, decreased in size in 43 percent, showed no change in 34 percent, and increased in 4 percent of patients. Menorrhagia improved in 95 percent of patients, with 46 percent having bleeding "reduced to normal." Similarly, symptoms of dysmenorrhea improved in 94 percent of patients with this symptom and were completely relieved in 51 percent. Fifteen of the 110 patients underwent hysterectomy during followup.

Methodological Issues

Symptom outcomes were poorly measured and reported in these trials. Nearly all studies required subjects to have symptoms associated with fibroids, principally menorrhagia and pelvic pain. When reporting the alleviation of symptoms, many studies either did not differentiate among various symptoms or did not report the number of women reporting certain types of symptoms. Because women may present with multiple symptoms, a combined-symptom measure would be preferable to enumerating the response for each symptom separately. None of the studies used any kind of systematic symptom inventory or disease-specific quality-of-life measure.

One article on oral contraceptives was subsequently withdrawn from publication without explanation (Friedman and Thomas, 1995). Without more explanation, it is unclear whether this retraction should affect the assessment of the validity of other publications by the same authors.

Summary

GnRH agonist treatment is associated with reduction of fibroid and uterine size, control of bleeding, and reduction of some symptoms; however, these drugs cause menopausal symptoms and, with long-term use, bone loss. Followup studies show that after treatment with GnRH agonists is stopped, there is regrowth of both the fibroid tumors and the uterus to near pretreatment size, which is often associated with the return of fibroid-related symptoms.

Concurrent treatment with GnRH agonists and low doses of progestins or estrogen-progestin combinations tends to ameliorate vasomotor menopausal symptoms (hot flashes) associated with GnRH agonists; however, this add-back treatment was associated with irregular bleeding in some studies. Androgen treatment following GnRH treatment may be effective at reducing regrowth of fibroids shrunk by GnRH treatment; also, GnRH agonist therapy may reduce regrowth of fibroids following myomectomy; however, the effect on fibroid-related symptomatology is less clear.

Used in the short term prior to myomectomy, GnRH agonists can correct anemia from menorrhagia (which may permit autologous blood donation), reduce blood loss during surgery, and allow use of transverse rather than midline incision or vaginal (or, conceivably, laparoscopic) rather than abdominal surgical approach due to reduction in uterine size. The clinical significance of statistically significant differences in estimated blood loss, operating time, or incision type is unclear, especially since long-term data from prospective studies of hysterectomy suggest that short-term complications or technical approach to surgery do not influence long-term outcomes. The effectiveness of using GnRH agonists to facilitate autologous blood donation for intra- or postoperative transfusion was not explicitly studied among patients with fibroids. Data on cost-effectiveness would be useful. There are two additional concerns regarding presurgical treatment with GnRH agonists:

1. The observation by some surgeons that fibroids are more difficult to separate from the uterus after GnRH agonist treatment is supported by data from hysterectomy specimens, which suggests that pretreatment with GnRH agonists obliterates the cleavage plane between myometrium and fibroid (Deligdisch, Hirschmann, and Altchek, 1997).

2. Pretreatment with GnRH may increase the postmyomectomy fibroid recurrence rate by shrinking some small fibroids so that they inadvertently are not removed at the time of myomectomy. This concern is supported by two studies reporting data on recurrence (as measured by ultrasound) (Fedele, Vercellini, Bianchi, et al., 1990; Friedman, Daly, Juneau-Norcross, et al., 1992).

Although most of the trials considered here were limited to women with symptomatic uterine fibroids, with symptoms usually defined as vaginal bleeding (menorrhagia), constant or cyclic pelvic pain, or constipation or urinary symptoms, the trials less often describe symptomatic outcomes other than uterine or fibroid size changes. Data from a limited number of trials, however, show a strong association between the uterine and fibroid size reduction and alleviation of symptoms.

Control groups on placebo or no treatment for prolonged periods demonstrate that patients with symptomatic fibroids for whom nothing is done (no treatment) remain stable in terms of symptoms or uterine and fibroid size. However, these studies were performed in women with mild symptoms. The prognosis for women with more severe fibroid-related symptoms or larger fibroids may be different.

Hormone replacement therapy with cyclic or noncyclic estrogen-progestin combinations appears to be ineffective in alleviating fibroid symptoms or fibroid growth. Progestins are ineffective alone, when used as add-back therapy during GnRH agonist treatment, or when used following GnRH agonist treatment to maintain fibroid control. However, progestins are effective at eliminating vasomotor symptoms of hot flashes associated with GnRH agonists.

Approach

We attempted to collect data on the costs of all the targeted treatments for fibroids, even though evidence for effectiveness was lacking for many of them. We first identified the types of data that would be needed to provide a comprehensive description and analysis of the costs associated with treatments for fibroids. Other considerations often considered as "costs" -- stresses on family life, effects on sexual function -- are more appropriately considered as quality-of-life measures, to be incorporated into the denominator of a cost-effectiveness ratio. Nonmedical costs include:

• Decreased productivity associated with bleeding or pain, either through actual time away from work or decreased ability to perform work because of symptoms.

• Decreased productivity associated with diagnostic and therapeutic interventions; time lost from work would likely affect other family members as well for some interventions.

• Transportation costs associated with provider visits.

• Child care expenses associated with provider visits.

• Use of nonmedical supplies, such as sanitary napkins.

Medical costs for the interventions considered in this report include the following:

For no treatment (for an asymptomatic patient):

• Costs associated with increased surveillance (more frequent pelvic examinations or radiological studies) after detection.

• Costs associated with treatment once uterine or fibroid size has reached a predefined threshold (we did not identify evidence that would permit estimation of an appropriate threshold).

• Costs associated with increased morbidity if the probability of success or the technical difficulty of treatment has increased; evidence supporting increased morbidity associated with increasing fibroid size is inconclusive (see under Question 1).

For oral contraceptives, progestins, and NSAIDS:

• Cost of medications.

• Cost of managing side effects.

• Cost of provider visits during followup.

• Cost of diagnostic tests during followup.

For GnRH agonists:

• Cost of medication.

• Cost for administration of injection for injectable forms.

• Costs of any medications used to manage side effects.

• If used as an adjunct for surgical procedures, any differences in costs between procedures done with and without adjunctive treatment.

For invasive therapies (embolization, myomectomy, hysterectomy):

• Cost of equipment and supplies used to perform the procedure.

• Fixed costs associated with use of radiology suite.

• Personnel costs.

• Medication costs for management of pain.

• Hospitalization costs for patients admitted after procedure.

• Cost of managing complications.

• Followup outpatient visits and testing, if indicated.

We were unable to identify a data source for estimating nonmedical costs associated with the management of asymptomatic fibroids. For estimating medical costs, we used several sources, including the following:

• 2000 "Red Book" of wholesale drug prices (Medical Economics Company. 2000).

• Published literature on hospital costs for surgical management of fibroids.

• Primary data from the Nationwide Inpatient Sample (NIS).

• Primary data from Duke University Medical Center.

Results

Medical Therapy

Costs for a 3-month course of medical therapy were estimated by:

• Identifying minimum and maximum average wholesale prices for medications (including generics if available) using the 2000 "Red Book" of pharmaceutical prices.

• Estimating monthly pill usage based on potential dosing regimens (e.g., up to 7 days of NSAID therapy, or 10 days of oral medroxyprogesterone acetate).

• Estimating monthly costs based on wholesale prices for a given number of dispensed units (e.g., if 100 pills were dispensed at a given price and 21 would be used during a month, the monthly cost was estimated as 0.21 x cost/100 pills).

Costs are given for 3 months to facilitate comparison with 3 months of GnRH agonist therapy, the maximum time GnRH would be given prior to surgical management.

Table 23. Average wholesale prices for medical treatments (more...)

Table 23. Average wholesale prices for medical treatments for symptomatic fibroids

Drug	Minimum price	Maximum price	3-month minimum	3-month maximum
NSAIDs
Ibuprofen
200 mg (100 pills)	$1.75	$6.18	$2.94	$15.57
400 mg (100 pills)	$17.02	$26.48	$14.30	$22.24
600 mg (100 pills)	$27.39	$38.69	$17.26	$24.37
800 mg (100 pills)	$26.21	$45.34	$16.51	$28.56
Naproxen
275 mg (30 pills)	$24.84	$38.20	$20.87	$32.09
550 mg (30 pills)	$39.21	$130.69	$16.47	$54.89
Mefanamic acid
250 mg (30 pills)	$42.08	$42.08	$35.35	$35.35
Oral contraceptives (6 months)
Progestin only	$185.09	$208.92	$92.55	$104.46
Combination	$168.36	$281.19	$84.18	$140.60
Progestins
Oral medroxyprogesterone acetate 10 mg (30 pills)	$19.87	$31.21	$59.61	$93.63
Depot medroxyprogesterone acetate	$48.10	$48.10	$48.10	$48.10
GnRH agonist
Depot leuprolide Acetate	$478.01	$518.64	$1,434.03	$1,555.92

Results are shown in Table 23. NSAIDs are the least expensive treatment. The figures in Table 23 represent prolonged monthly courses of 7 days; costs for women who respond to shorter courses would be lower. Long-acting medroxyprogesterone acetate is next; unfortunately, there are no data on its effectiveness in women with fibroids. Oral progestins and oral contraceptives are next, at two to five times the cost of NSAIDs. GnRH agonists are the most expensive.

Estimation of cost-effectiveness would require better data on effectiveness, as well as on costs. One analysis of the use of adjunctive GnRH agonists prior to surgery suggested that their use might be cost-effective because of decreased complication rates and increased ability to perform less expensive vaginal hysterectomy (Bradham, Stovall, and Thompson, 1995). Replication of this analysis using more recently published effectiveness and cost data would be worthwhile.

Myomectomy

Table 24. Mean total hospital costs for myomectomy, (more...)

Table 24. Mean total hospital costs for myomectomy, by age

Age	Mean total cost (n = 239)
< 25	$4,340 ± 663
25-29	$4,796 ± 694
30-34	$5,105 ± 921
35-39	$5,308 ± 2,066
40-44	$5,333 ± 1,842
45-49	$5,684 ± 1,962
> 50	$4,490 ± 704

Source: Duke University Medical Center, 1992-1998

Estimated mean charges for women 18-44 years of age undergoing myomectomy in the NIS for 1997 were $8,884. Mean total costs for women undergoing the procedure at Duke University Medical Center from 1993 through 1998 (normalized to 1999 dollars) by age are shown in Table 24. The total mean Duke cost was $5,171, consistent with a 50-60 percent cost-to-charge ratio compared with the NIS data. Of these costs, 30.4 percent were attributable to inpatient unit costs, and 54.0 percent were attributable to surgical services.

Hysterectomy

Table 25. Mean total hospital costs for hysterectomy, (more...)

Table 25. Mean total hospital costs for hysterectomy, by age

Age	Mean costs (n = 753)
< 25	--
25-29	$4,490 ± 704
30-34	$5,309 ± 1,413
35-39	$5,577 ± 1,908
40-44	$5,656 ±1,809
45-49	$6,348 ± 2,627
> 50	$6,222 ± 2,469

Source: Duke University Medical Center, 1992-1998

Mean Duke costs by age group are shown in Table 25 (all hysterectomies combined). Overall mean cost was $5,961, with 32.6 percent attributable to inpatient unit costs and 49.0 percent attributable to surgical services.

Methodological Issues

Data are not available to estimate the overall economic costs to society of uterine fibroids. Data also are not available on the costs of outpatient management, other than estimates of the wholesale prices of drugs (which clearly do not reflect actual prices paid by providers or patients). Although relatively detailed data on inpatient resource utilization are available, additional work on clinical factors that predict resource utilization needs to be done.

Even with better data on costs, cost-effectiveness analysis cannot be performed without better data on effectiveness. Even if NSAIDs are the least expensive method of treatment, their cost-effectiveness ratio will be quite high if their effectiveness (the denominator in the ratio) is low.

Summary

Data on nonmedical costs and outpatient costs of managing patients with fibroids are not readily available from easily accessed sources. Although data on inpatient costs are somewhat more detailed, most administrative data sources do not provide sufficient clinical detail to allow comparison between procedures. At one academic medical center, total mean inpatient costs for abdominal myomectomy were approximately $800 less than total costs for hysterectomy.

Approach

We approached the question of different benefits and risks of strategies for managing fibroids among subpopulations by focusing on four separate topics:

• Racial and ethnic differences in epidemiology and outcomes.

• Effects of age, especially menopausal status, on epidemiology and outcomes.

• Effects of fibroids on pregnancy outcomes and complications.

• Effects of fibroids on fertility.

Articles that met our screening criteria were searched and subcategorized according to the above topics.

Race and Ethnicity

Ten articles were identified initially as being potentially informative in answering questions about whether racial distinctions exist in the risks and benefits of fibroid management. After closer inspection, only four of these actually proved to be useful for this analysis (Hillis, Marchbanks, and Peterson, 1996; Kjerulff, Guzinski, Langenberg, et al., 1993; Kjerulff, Langenberg, Seidman, et al., 1996; Marshall, Spiegelman, Barbieri, et al., 1997). All four articles focused on a comparison between black and white populations. Two included other racial groups, including Hispanics and Asians, with results statistically indistinguishable from the white population with whom they were compared (Hillis, Marchbanks, and Peterson, 1996; Marshall, Spiegelman, Barbieri, et al., 1997). In both studies, "white" was used as the reference group, with no separate analyses looking at differences within other racial groups. The rest of this discussion will focus on the distinctions between blacks and whites unless otherwise specified.

Management

Although the time from diagnosis of uterine fibroids to hysterectomy has been shown to be longer for blacks than for whites (3.9 ± 5.0 versus 2.8 ± 3.6 years) (Marshall, Spiegelman, Barbieri, et al., 1997), black women overall undergo hysterectomy at a younger age than white women due to their earlier average age at diagnosis. In one prospective cohort, the average age at hysterectomy for blacks was 41.7 ± 6.1 versus 44.6 ± 5.8 for whites (Kjerulff, Langenberg, Seidman, et al., 1996). In one large epidemiological study of more than 53,000 women, uterine fibroids accounted for 65.4 percent of the hysterectomies in black women versus only 28.5 percent in white women (Kjerulff, Guzinski, Langenberg, et al., 1993). Moreover, at the time of hysterectomy, black women have been shown on average to have both more myomas and a larger uterine size (Hillis, Marchbanks, and Peterson, 1996; Kjerulff, Langenberg, Seidman, et al., 1996). Kjerulff and colleagues found that 30 percent of black women who underwent hysterectomy had uteri greater than 500 grams versus only 15 percent of white women (Kjerulff, Langenberg, Seidman, et al., 1996). This difference has potentially important implications because Hillis and colleagues demonstrated that women undergoing hysterectomy with a uterus of > 500 grams were 1.6 (1.0-4.0) times more likely to develop operative or postoperative complications, particularly cuff cellulitis (relative risk [RR], 2.8 [95% confidence interval (CI), 1.3-6.2]) and transfusion (RR, 2.4 [95% CI, 1.3-4.3]).

At the time just prior to hysterectomy, Kjerulff and colleagues found that black women were more likely to be anemic than white women (56 percent vs. 38 percent) and more likely to complain of severe pelvic pain (59 percent vs. 41 percent ) (Kjerulff, Langenberg, Seidman, et al., 1996). There was no evaluation of postoperative symptomatology in this study, and none of the other studies evaluated symptomatology along racial lines.

We identified no studies that evaluated racial differences in response to, or provision of, medical therapies for uterine fibroids, such as NSAIDs, GnRH agonists, or other hormonal preparations. Similarly, there are no published data evaluating racial distinctions concerning myomectomy, uterine artery embolization, or surgical procedures other than hysterectomy.

Figure 2. Age- and race-specific incidence of myomectomy, 1997, (more...)

   Figure 2. Age- and race-specific incidence of myomectomy, 1997, based on NIS and U.S. Census Bureau estimates

Figure 3. Age- and race-specific incidence (more...)

   Figure 3. Age- and race-specific incidence of hysterectomy for fibroids, 1997, based on NIS and U.S. Census Bureau estimates

Figure 4. Estimated cumulative incidence of (more...)

   Figure 4. Estimated cumulative incidence of myomectomy based on 1997 NIS and U.S. Census Bureau data

Note: Cumulative incidence may be overestimated, since individual women may undergo the procedure on more than one occasion.

Figure 5. Estimated cumulative incidence of hysterectomy (more...)

   Figure 5. Estimated cumulative incidence of hysterectomy for fibroids based on NIS and U.S. Census Bureau data

Preliminary analysis of data from the 1997 Nationwide Inpatient Sample (NIS, 1997) shows the following: 45,740 women were identified as undergoing either myomectomy (n = 7,749) or hysterectomy (n = 37,991) for fibroids. Figures 2 through 5 show the estimated age- and race-specific incidences of myomectomy and hysterectomy based on 1997 NIS and U.S. Census Bureau estimates, as well as the projected race-specific cumulative incidences of each procedure based on these data. Bivariate analysis revealed that, at the time of their procedure, 45 percent of black women were younger than 40 versus 35 percent of Hispanic women and only 27 percent of white women (p < 0.0001). Data on number of prior children, contraceptive use, or history of sterilization are not available within the NIS data. Hispanics were more likely to have Medicaid/Medicare than their black and white counterparts: the figures were 23 percent, 17 percent, and 17 percent respectively (p < 0.0001). Blacks were more likely than either whites or Hispanics to earn $25,000 or less (41 percent vs. 29 percent vs. 26 percent) and less likely to earn more than $35,000 (24 percent vs. 31 percent vs. 36 percent) (p < 0.0001). Blacks were more likely than either whites or Hispanics to earn $25,000 or less (41 percent vs. 29 percent vs. 26 percent) and less likely to earn more than $35,000 (24 percent vs. 31 percent vs. 36 percent (p < 0.0001).

Thirteen percent of white women were treated with myomectomy versus 22 percent of black women and 17 percent of Hispanic women. Results for both the unadjusted and adjusted models were similar and revealed that, when undergoing a procedure for fibroids, black women are 1.6 (95% CI, 1.5-1.7) times more likely to receive a myomectomy than white women. Similarly, Hispanic women are 1.3 (95% CI, 1.2-1.5) times more likely to receive a myomectomy than white women. These findings were consistent for every age group of women. This suggests that, at the least, minority women are not less likely to be offered conservative surgical therapy than white women. There also may be cultural differences that lead some minority women to be more likely to request conservative therapy.

Preliminary analysis of the primary data abstraction of 239 patients undergoing myomectomy at Duke University Medical Center also shows some interesting findings. Black women were more than twice as likely as white women to have an in-hospital complication (odds ratio [OR], 2.48) or transfusion (OR, 2.2), but this increased risk was eliminated after adjusting for uterine size and number of fibroids (adjusted OR for complications 1.36 [95% CI, 0.56-3.15]; adjusted OR for transfusion 0.9 [95% CI, 0.27-2.76]). Other factors, such as body mass index (BMI) or insurance status, were not independent predictors of complications. This suggests that, for myomectomy, an increased complication rate in black women is largely attributable to differences in the underlying disease pathology. Currently, we are analyzing data on hysterectomy from the same time period to see if a similar pattern is observed.

Age/Menopausal Status

There were 49 articles identified as being potentially informative in evaluating how the risks and benefits of fibroid management differ by menopausal status. Of these, 34 were found useful in providing information on this topic.

Pregnancy Outcomes

Pregnancy After Invasive Uterus-Conserving Therapy

The studies looking at pregnancy complications postmyomectomy all have been small and descriptive in nature. Loss of clinically identified pregnancies in women postmyomectomy has been reported to range from 12-29 percent (Dubuisson, Chapron, and Levy, 1996; Li, Mortimer, and Cooke, 1999; Sudik, Husch, Steller, et al., 1996). One small case series (n = 46) with 41 months followup noted that "habitual aborters" had a 10 percent live birth rate premyomectomy versus an 87 percent live birth rate postoperatively (Vercellini, Maddalena, De Giorgi, et al., 1999). Several reviewers of this report pointed out case reports of uterine rupture in early pregnancy after laparoscopic myomectomy, but as discussed above, estimation of true risks based on case reports alone is extremely difficult. Although it has been widely believed that women should undergo elective c-section for delivery after a myomectomy, especially when the uterine cavity has been entered, data to support this recommendation are limited. A study by Garnet found that 55 percent of all ruptures during pregnancy had a previous uterine scar; however, fewer than 5 percent of these women had had a myomectomy. Moreover, uterine rupture occurred in less than 1 percent of pregnancies with prior uterine scars, with two-thirds occurring prior to labor onset (Garnet, 1964), suggesting that planned elective cesarean delivery will not eliminate the risk of uterine rupture in women with uterine scars.

The majority of reported series of uterine artery embolization include one or two unplanned, uneventful pregnancies after the procedure. We identified one case series of 12 pregnancies after uterine artery embolization in nine women (Ravina, Vigneron, Aymard, et al., 2000). Although this paper did not meet our usual inclusion criteria, we discuss it here because of intense interest in pregnancy-related outcomes of this procedure. There were no recurrences of fibroids during any of the pregnancies. Five of the 12 pregnancies resulted in early pregnancy loss (all in women over 40). There were two preterm deliveries at 28 weeks (an AIDS patient with sepsis) and 35 weeks (preeclampsia with twins) and five term deliveries. Four deliveries were by cesarean, and three were vaginal. These numbers are too small to allow meaningful conclusions. The relatively advanced age of the women in this study (mean 36.5, with 5 of 9 patients at least 40 years old) makes them a particularly high-risk group for adverse pregnancy outcomes.

Fertility

Approach

The ideal approach to this question would be to compare prospectively collected long-term results in a group of women undergoing surgical management of fibroids with a similar group of women with fibroids who underwent alternative therapies. Unfortunately, such data do not exist. Even without a control group, published prospective studies of women undergoing hysterectomy have, at most, 2 years of followup data, and few report results separately for women undergoing hysterectomy for fibroids.

Because it is performed so frequently, there is a body of literature addressing some of the long-term effects of hysterectomy, although most of these studies have significant methodological limitations, which are described below. Unfortunately, few if any of these studies stratify results based on the indication for the procedure or comorbidities present at the time of surgery. This could result in significant biases. For example, women undergoing hysterectomy primarily for symptoms of prolapse and those who undergo concurrent urinary tract or vaginal surgery at the time of hysterectomy might well be more likely to develop urinary incontinence or pelvic floor dysfunction than women undergoing hysterectomy for bleeding or pain related to fibroids.

Despite these limitations, we elected to review the literature on the effects on the aging process of hysterectomy performed for any benign indication. Our goal was to provide some information for patients and providers to consider when making decisions about therapy and to identify methodological limitations and areas for future research targeted specifically to women with fibroids.

Specifically, after discussion with AHRQ and the technical advisory panel, we attempted to identify studies that addressed the effect of procedures used in the surgical management of fibroids on ovarian function, sexual function, and pelvic floor function, as well as studies that addressed the benefits and risks of prophylactic oophorectomy at the time of hysterectomy for benign disease.

Ovarian Function

Sexual Function after Hysterectomy

We identified 31 studies that address the effect of hysterectomy in general on sexual functioning, but only one of these (Carlson, Miller, and Fowler, 1994a) provided information that specifically addressed the question of sexual function in women who underwent hysterectomy for uterine fibroids. Furthermore, we were unable to identify any studies that addressed the effect of other invasive therapies for uterine fibroids, such as myomectomy or uterine artery embolization, on sexual function.

Pelvic Floor Dysfunction

For this question, we divided the spectrum of pelvic floor dysfunction into three broad categories: lower urinary tract dysfunction; dysfunction of the colon, rectum, and anus; and pelvic organ prolapse. Forty-one articles were identified that examined the effect of hysterectomy in general on disorders of the pelvic floor. We identified only one article that contained information specifically addressing hysterectomy for fibroids and subsequent pelvic floor dysfunction. Two additional articles were identified that had information on conservative surgery for uterine fibroids and urinary symptoms.

Prophylactic Oophorectomy

Model Structure and States

Figure 6. Schematic of fibroids model

   Figure 6. Schematic of fibroids model

Figure 6

• Asymptomatic fibroids (prior to any intervention).

• Symptomatic fibroids, no side effects/complications of therapy.

• Symptomatic fibroids, side effects/complications of therapy.

• Improved symptoms, no side effects/complications.

• Improved symptoms, side effects/complications.

• No symptoms, no side effects/complications (after an intervention that successfully resolves symptoms and does not result in any side effects or complications).

• No symptoms, side effects/complications.

• Uncomplicated pregnancy (no complications attributable to fibroids).

• Complicated pregnancy (complications attributable to fibroids).

• Menopause.

• Death.

These states were chosen based on our clinical experience with patients with fibroids and input from the advisory panel. The goal was to allow the model to reflect the history of the condition and its treatment in sufficient detail to permit estimates of the rate and timing of occurrence of specific outcomes of interest. This detail allows explicit presentation of different outcomes for different treatments -- benefits and risks that occur at different rates or with different timing.

Assumptions

We made several assumptions in constructing the model to make the structure more tractable and to permit estimation of model parameters from available data.

• The likelihood of symptomatic improvement is independent of the probability of side effects or complications. Although this is likely to be true for some symptomatic states (e.g., those related to bleeding), it may not be for others (e.g., those related to pain). For infertility, there are some complications (adhesions, necessity for hysterectomy) that may actually decrease the probability of success. However, given that we were unable to identify any study that reported data using similar groupings, this assumption seems reasonable. For hysterectomy, at least, data from the Maryland Women's Health Study suggest that the likelihood of symptomatic improvement is not related to the occurrence or severity of perioperative complications (Kjerulff, Langenberg, Rhodes, et al., 2000).

• Symptoms resolve with the onset of menopause. This assumption simplifies the model. However, if hormone replacement therapy is used, the assumption may be less valid. The model can be readily modified to allow continuation of symptomatic states into menopause, with appropriate adjustments in probabilities as data become available. For women who are perimenopausal, we assumed that symptoms will continue. Bleeding problems during the perimenopausal phase were common in the Massachusetts Women's Study (McKinlay, Brambilla, and Posner, 1992), so it seems likely that abnormal bleeding related to fibroids would persist during this phase.

• For simplicity, the model considers a "pregnancy" to occur when it has reached 24 weeks. Thus, "pregnancy" excludes failure to conceive, habitual abortion, or early second trimester loss. Again, because the likelihood of these outcomes, the associated costs, and their quality-of-life impact may well differ, the model will need to be modified to incorporate these outcomes specifically for those patients with concerns about future pregnancies.

Operation of the Model

The operation of the model is illustrated by the following description of a simulation run for a patient with symptomatic fibroids:

• The patient's age, race, and particular symptoms or problem (bleeding, dysmenorrhea, infertility, etc.) are specified at the start of the simulation. Depending on the treatments under consideration, other factors that might influence the likelihood of particular outcomes (other causes of infertility, comorbid conditions, severity of disease) would also be specified.

• All future health states are tracked for discrete cycles (the default cycle length in this model is 1 year, although shorter lengths might be appropriate in certain situations). At the beginning of each cycle the patient may either become menopausal or die of other causes; these probabilities are age-dependent and can be adjusted for race.

• If future pregnancy is not a consideration, the age-dependent probability of pregnancy (modifiable by treatment or underlying morbidity) can be "switched off."

• Patients who elect no treatment have a probability of spontaneous resolution of symptoms or a lessening of the degree to which the symptoms are bothersome; for the purposes of the model, these two are equivalent. Alternatively, symptoms can persist or worsen.

• Patients who elect treatment have a treatment-specific probability of improved symptoms (which can be further divided into complete and partial resolution), as well as treatment-specific probabilities of side effects and/or complications.

Note that because the length of the simulation can be varied, long-term data, if available, can be incorporated. For example, patients undergoing surgical management who experience short-term complications may transition to a "no side effect/complication" branch after one or two cycles. Similarly, recurrence or need for additional treatment after conservative therapy also can be incorporated.

Model Transition Probabilities

Table 26. Probabilities used for estimating outcomes (more...)

Table 26. Probabilities used for estimating outcomes of hysterectomy, myomectomy, and no treatment

Probability	Estimate	Range
All options
All cause mortality	Age-specific, from life-tables
Menopause	Age-specific, derived above
No treatment option
Spontaneous resolution	30% at 1 year	20-40%
Need for hysterectomy	10% at 1 year	10-20%
Myomectomy option
Symptomatic improvement	90%	60-95%
Short-term complications	Equivalent to hysterectomy
Recurrence leading to hysterectomy	3% at 2 years	3-10%
Long-term complications	0%	0%
Hysterectomy option
Symptomatic improvement	90%	60-95%
Short-term complications	Equivalent to myomectomy
Long-term complications	12% at 2 years	0-15%

Natural History Parameters

Mortality

Age-specific mortality rates were obtained from 1997 U.S. life-tables (Anderson, 1999). The probability (p) of an event occurring in a time interval (d) based on an event rate (r) is calculated using the standard formula: p = 1 − e^(−r)x(d).

Menopause

The age-specific rate of becoming amenorrheic secondary to natural menopause was estimated from several sources. For women aged 30 to 39, Coulam and colleagues reported a rate of 76/100,000 per year derived from a cohort study in Minnesota, resulting in a prevalence of 1 percent by age 40 (Coulam, Adamson, and Annegers, 1986). This rate was converted to an annual transition probability as above.

Probabilities for women aged 45 through 55 were estimated from cross-sectional and prospective data from the Massachusetts Women's Study (McKinlay, Brambilla, and Posner, 1992). McKinlay and colleagues reported on the transition from premenopause through perimenopause to postmenopause.

Annual transition probabilities from pre- and perimenopause to postmenopause were derived by estimating the age-specific prevalence of postmenopause from figures in McKinlay and colleagues (McKinlay, Brambilla, and Posner, 1992) and calculating the age-specific transition rates required to achieve those prevalences for a cohort beginning at age 45.

Probabilities between age 40 and 45 were estimated by fitting an exponential curve to the observed data:

Proportion postmenopausal = 0.0115 * e^{0.6989 x}, where x = number of years after age 40.

Table 27. Estimated age-specific transition rates from pre- (more...)

Table 27. Estimated age-specific transition rates from pre- and perimenopausal to postmenopausal state

Age	Proportion post-menopausal	Transition rate/year	Transition probability
39	0.0089	1/1000	0.0008
40	0.010	4/1000	0.0040
41	0.014	5/1000	0.0050
42	0.019	7/1000	0.0070
43	0.026	7.2/1000	0.0072
44	0.039	32/1000	0.0315
45	0.07	32.2/1000	0.0317
46	0.1	55/1000	0.0535
47	0.15	59/1000	0.0573
48	0.2	94/1000	0.0897
49	0.275	103/1000	0.0979
50	0.35	215/1000	0.1935
51	0.49	216/1000	0.1943
52	0.6	300/1000	0.2592
53	0.72	300/1000	0.2592
54	0.8	286/1000	0.2487
55	0.85	250/1000	0.2212

Note: Based on data from McKinlay, Brambilla, and Posner, 1992

The resulting transition probabilities used in the simulation are shown in Table 27.

Natural history of symptomatic fibroids

Relevant probabilities for a woman with symptomatic fibroids who does not undergo any further therapy include (1) spontaneous resolution of symptoms, (2) no change in symptoms, and (3) worsening of symptoms resulting in the need for surgical therapy. The likelihood of spontaneous resolution of symptoms was obtained from the nonsurgical cohort in the Maine Women's Health Study (Carlson, Miller, and Fowler, 1994b). Unfortunately, results in this study are reported for all nonsurgical treatments, including hormonal therapy and NSAIDs rather than for individual treatment options. Because 68 percent of women with fibroids received no treatment, we have assumed that the results for no treatment alone are not substantially altered by the inclusion of women receiving medical therapy. At 1 year after initiation of the cohort, the proportion of women with fibroids who reported that their symptoms were not bothersome increased from 47 percent to 63 percent. This is equivalent to a resolution rate of 30 percent per year.

Carlson and colleagues (Carlson, Miller, and Fowler, 1994a) reported a 23 percent incidence of hysterectomy at the end of 1 year, with the rates not differing between indications. Weber and colleagues reported a 22 percent incidence over 29 months (Weber, Mitchinson, Gidwani, et al., 1997). These two incidences result in a probability range of 9.5 to18 percent. In the base case, we use the lower end.

Assigning Values to Outcomes

In this example, we again simplify by assigning a value of 1 for complete relief or entering menopause and 0 for all other states (persistent symptoms, long-term complications, or death). Obviously, values for these states will differ between patients, and temporary states (such as short-term complications) or patient preferences for outcome process (a strong preference to avoid hysterectomy, for example) play important roles. However, this simplified system helps illustrate how the various probabilities in the model affect outcomes.

Outcomes

The primary outcome of interest is relief of symptoms. In this illustrative example, response to treatment is dichotomous, either complete relief or no relief. Secondary outcomes include either immediate complications or development of new symptoms after treatment.

Analytic Horizon

Because the longest published followup in prospective cohort studies of hysterectomy patients is 2 years, the analytic horizon for this analysis is 2 years.

Sensitivity Analysis

To assess the impact of uncertainty in parameter estimates, we performed one-way sensitivity analysis. Outcomes were calculated for the range of plausible values.

Software

The tree was constructed and all calculations performed using DATA 3.5 (TreeAge, Williamstown, MA).

Base Case

Using the base case assumptions, the probabilities of complete relief or menopause, without development of long-term complications, for a 35-year-old woman choosing between these options are 35.8 percent for no treatment, 83.4 percent for hysterectomy, and 89.9 percent for myomectomy.

Sensitivity Analyses

We performed sensitivity analysis on all parameters by varying the values across the ranges described above, both individually and in two- and three-way combinations. We illustrate the effect of uncertainty surrounding four specific parameters, all of which are related to key questions addressed in this report, in the figures that follow. In all figures, the vertical axis represents the "expected value" of the decision -- the probability of a favorable outcome multiplied by the value (1.0, in this case) of a favorable outcome, plus the probability of an unfavorable outcome multiplied by the value of an unfavorable outcome (0).

Long-Term Complications

Figure 7. Sensitivity analysis on probability of long-term complications (more...)

   Figure 7. Sensitivity analysis on probability of long-term complications after hysterectomy

The probability of a favorable outcome is higher with hysterectomy if the probability of a long-term complication is less than 5.2 percent (see Figure 7).

Recurrence after Myomectomy

Figure 8. Sensitivity analysis on probability of recurrence after (more...)

   Figure 8. Sensitivity analysis on probability of recurrence after myomectomy

Recurrence risk did not substantially affect the probabilities (see Figure 8).

Spontaneous Resolution of Symptoms

Figure 9. Sensitivity analysis on probability of improvement with (more...)

   Figure 9. Sensitivity analysis on probability of improvement with no treatment

As the likelihood of spontaneous resolution increased, the likelihood of a favorable outcome also increased, although this probability always remained less than for myomectomy or hysterectomy (see Figure 9).

Patient Age

Figure 10. Sensitivity analysis on patient's age

   Figure 10. Sensitivity analysis on patient's age

Because patient age strongly affects the likelihood of undergoing menopause, the probabilities of a favorable outcome of each strategy also were affected by patient age (see Figure 10).

The likelihood of a favorable outcome with no therapy increases substantially with increasing age. This likelihood also increases slightly for the surgical options because of the increased probability of menopause in those women who do not respond favorably to the initial therapy.

What are the risks and benefits of hysterectomy and myomectomy in the treatment of symptomatic and asymptomatic fibroids?

There is no evidence documenting any benefit for the use of either hysterectomy or myomectomy for management of asymptomatic fibroids. Each of these procedures has definite risks. We did not identify any randomized trials directly comparing hysterectomy with myomectomy in women with symptomatic fibroids.

The majority of reviewed studies of abdominal and laparoscopic myomectomy did not report on the effects of the procedure on relieving symptoms. Studies of hysteroscopic myomectomy did report symptomatic relief, but followup and methods for determining symptom severity were variable. Results of hysteroscopic myomectomy appear to be related to fibroid anatomy, with success decreasing as a greater percentage of the fibroid involves the myometrium. Blood loss resulting in blood transfusion was the most commonly reported short-term complication of all myomectomy procedures, ranging from 2-10 percent. Hysteroscopic procedures carried risks of uterine perforation and fluid/electrolyte disturbances.

Three prospective studies of women undergoing hysterectomy reported overall favorable outcomes at followup of up to 2 years; results for women with fibroids appeared to be consistent with results for the overall cohort. Up to 12 percent of women reported new symptoms after hysterectomy, with menopausal symptoms being most common.

Unfortunately, because of the lack of data that would allow comparison between the baseline characteristics of patients undergoing each procedure, no inferences can be drawn about relative short- or long-term risks and benefits. Although some retrospective multivariate analyses suggest that much of an observed increased risk of hysterectomy is because of larger uterine size, this needs to be confirmed in larger studies.

What are the risks associated with single versus multiple myomectomies?

Women with a single clinically apparent fibroid who have a surgical removal may be less likely to experience short-term complications or long-term recurrence, and they may have higher pregnancy rates than women with multiple fibroids who undergo surgical management. However, variations in reporting methodology prevent quantitative synthesis of these findings. In addition, it is unclear whether these variations in outcomes are related to differences in surgical success rates or underlying patient biology.

Who are appropriate candidates for each procedure?

Women who desire to retain childbearing potential obviously are not candidates for hysterectomy. There is insufficient evidence to draw any conclusions about appropriate candidates for other procedures or treatments based on patient characteristics (beyond issues such as overall risk of complications from surgery because of medical comorbidity, contraindications to particular medications, and so forth).

What is the incidence of need for additional treatment after myomectomy or other uterus-sparing interventions?

It is clear that some women do develop symptomatic fibroids after treatment with medical therapy, uterine artery embolization, or myomectomy. Reported cumulative recurrence rates after myomectomy range up to 50 percent at 5 years, with up to 14 percent of patients subsequently undergoing hysterectomy. Increasing preoperative uterine size, increasing number of fibroids, increasing fibroid penetration into the myometrium, and residual tissue at the completion of the procedure were all associated with recurrence risk in at least two series. However, inconsistency in reporting definitions of recurrence, loss to followup, length of followup, use of other adjunctive treatments, and patient characteristics that might affect recurrence preclude a more precise estimation of incidence. In addition, it is not at all clear that "recurrent" fibroids after conservative therapy do not simply represent de novo development of additional fibroids; in this case, the association between "recurrence" and number of preoperative fibroids may represent an individual patient's propensity to develop myomas or the results of surgically induced changes in the uterus that actually facilitate fibroid growth.

Does additional treatment result in significantly increased morbidity?

Although there are potential mechanisms by which additional treatment might result in significantly increased morbidity, we were unable to identify any studies that provided information that would allow us to address this question. The trade-offs in terms of risks and benefits of immediate definitive therapy (i.e., hysterectomy) compared with conservative therapy, with a possible need for definitive therapy at a later date, are unclear. If additional treatment does result in increased morbidity (an increased relative risk), patients and clinicians need to take into account the actual probability of needing additional treatment (for which there are no data) in order to estimate an individual patient's overall absolute morbidity risk.

What are the risks and benefits of nonsurgical treatment?

We found surprisingly little literature on the benefits and risks of commonly used medical treatments for fibroids, such as nonsteroidal agents, progestins, and oral contraceptives. Data on long-term outcomes of GnRH agonist therapy also were lacking, although there is consistent evidence from randomized trials that preoperative use of GnRH agonist therapy reduces estimated blood loss and operating room time and may allow changes in surgical technique. The short- and long-term clinical significance of these findings is unclear.

What are the costs associated with effective surgical and nonsurgical treatments?

We found no data on nonmedical costs or outpatient costs other than prescription drug costs for diagnosis and management of symptomatic fibroids. Nonsteroidals are the least expensive of the medical treatments, followed by progestins and oral contraceptives. Drug costs for a 3-month course of GnRH therapy are in the range of $1,500. Data on hospital costs suggest that abdominal myomectomy is somewhat less expensive than hysterectomy, with a mean cost differential of approximately $800. Determining the relative cost-effectiveness of different treatments depends on data on effectiveness, which are lacking.

Do risks and benefits differ for women according to race, ethnicity, age, and interest in future childbearing?

What are the effects of surgical management of uterine fibroids, especially hysterectomy, on the aging process?

We found no evidence on the effects of myomectomy on the aging process. We found scant evidence on the effects of hysterectomy specifically performed for fibroids on the aging process. Hysterectomy in general does appear to affect ovarian function, even without removal of the ovaries. It does not appear to adversely affect sexual function in the short term in most women, and it may improve sexual function in many women with preoperative sexual dysfunction. Data on the effects of hysterectomy on pelvic floor dysfunction are scanty. Removal of both ovaries at the time of hysterectomy increased the likelihood of development of new symptoms or failure to improve from preoperative symptoms in one large prospective study.

Study Design

Given that there is no Food and Drug Administration (FDA) requirement for trials before the introduction of surgical procedures, strong patient and provider beliefs about the risks and benefits of surgery, and the methodological difficulties of performing randomized trials of surgical procedures, it is not surprising that randomized trials of surgical treatments for fibroids are exceedingly rare. However, given the frequency with which fibroids are diagnosed and the almost universal recommendation for a trial of medical management prior to surgical intervention (Broder, Kanouse, Mittman, et al., 2000), we expected to find better evidence supporting the use of these medical therapies. Even the best-designed cohort study reporting on outcomes of nonsurgical treatment (Carlson, Miller, and Fowler, 1994b) presents combined results for no therapy, hormones, and nonsteroidals, precluding estimation of a treatment effect. Until well-designed trials of the effectiveness of nonsurgical treatments are performed, insisting on a trial of medical management prior to definitive surgical therapy is somewhat difficult to justify on effectiveness grounds alone, especially when the quality of the evidence supporting the effectiveness of hysterectomy is superior to that supporting most medical treatments. If the major symptom is bleeding, it should be acknowledged that the best available data show that hysterectomy has nearly 100 percent effectiveness and provides definitive therapy.

Comparability of Subjects

Although randomized trials, including trials of surgical therapies, would be ideal, other study designs are likely to continue to be reported. The ability to synthesize data from these studies would be significantly enhanced by common standards for reporting important characteristics of the patients and treatment. This is especially true for factors likely to affect short-term outcomes (e.g., complications). At a bare minimum, data on uterine size, number of fibroids, size of fibroids, and location of fibroids should be provided. These data should be reported in a way that would facilitate comparison between studies and use in multivariate methods. For example, reporting of distributions of number of fibroids may be a more useful way of characterizing the patient population than reporting means and standard deviations, especially since noninteger numbers are meaningless.

Comparability of Symptom Measures

The ability to compare studies would be significantly enhanced by use of common measures, preferably measures that have been validated in appropriate populations. This specifically includes comparability in definition and ascertainment of persistence or recurrence after conservative therapy. Similarly, uniform timing of measurement before and after treatment would allow direct comparison of true rates of improvement, persistence, or recurrence by facilitating calculation of these events over time.

Duration of Followup

Even the best-designed and implemented studies of hysterectomy outcomes are limited to a 2-year followup. Longer term prospective studies, with data reported specifically by indication, are needed.

Literature Search

We used standard methods for identifying, reviewing, and abstracting published studies focused on management of uterine fibroids. Our criteria for study inclusion were notably less strict than those used in other systematic reviews, such as those conducted by the Cochrane Collaboration. We included a variety of nonrandomized study designs, including case series and studies with as few as 20 patients. We did not search the literature prior to 1975, primarily because we assumed that surgical technique (such as use of prophylactic antibiotics) and reporting standards had changed sufficiently to prevent generalization of results. We also limited our search to articles published in English, primarily for reasons of convenience and resources. It is possible that including older studies or studies published in other languages would have identified additional evidence that would have substantially changed our conclusions. This may be especially true for "alternative" therapies, such as herbal medications or for studies performed in populations where extensive emigration to the United States is a factor. Another limitation of our exclusion criteria is that rare but severe complications of treatments may have been overlooked because they were reported as case reports or small case series.

Grading Articles

We used a nonstandard method to grade the articles we reviewed. However, we believe that the rationale for each criterion is reasonable, and that the operational definitions are clear and reproducible. In addition, we used the grading criteria primarily to provide additional detail to other researchers; we did not establish a "threshold" for including articles or use these criteria to weight the results of a quantitative analysis.

Other Data Sources

We used two additional data sources for this report, the Nationwide Inpatient Sample (NIS) and primary chart abstraction from Duke University Medical Center. The NIS, like most administrative databases, is limited by a lack of clinically relevant detail that may affect short-term outcomes (Myers and Steege, 1999). In addition, economic data are based on patient charges rather than costs.

The Duke data, like most chart abstractions, are limited by inconsistency in clinical charting and missing data; data from an academic medical center also may not be generalizable to other settings. Sample size, while larger than previously published studies, is also somewhat small, especially for performing multivariate analyses.

Both data sets lack data on patient symptoms, response to treatment, and long-term outcomes.

We were unable to identify any readily accessible additional data sets that would supplement these or allow estimation of nonmedical costs or the costs of outpatient care, although it is possible that such data sets exist.

Decision Model

Like all models, ours has simplifying assumptions. It is possible that these assumptions invalidate the model or would prohibit its use in certain populations. It is also possible that our estimation of clinically important states for use in the model are not valid or do not represent states or concepts important to patients and clinicians. However, the main limitation to the use of the model is the lack of input data available for estimating transition probabilities and outcomes. Given the quality of the literature available for validation of the model, we clearly do not intend for it to be used in its current state to make treatment or policy decisions.