The REST Investigators, Abstract
Background
The efficacy and safety of uterine-artery embolization, as compared
with standard surgical methods, for the treatment of symptomatic
uterine fibroids remain uncertain.
Methods
We conducted a randomized trial comparing uterine-artery
embolization and surgery in women with symptomatic uterine fibroids.
The primary outcome was quality of life at 1 year of follow-up, as
measured by the Medical Outcomes Study 36-Item Short-Form General
Health Survey (SF-36).
Results
Patients were randomly assigned in a 2:1 ratio to undergo either
uterine-artery embolization or surgery, with 106 patients undergoing
embolization and 51 undergoing surgery (43 hysterectomies and 8
myomectomies). There were no significant differences between groups in
any of the eight components of the SF-36 scores at 1 year. The
embolization group had a shorter median duration of hospitalization
than the surgical group (1 day vs. 5 days, P<0.001) and a shorter
time before returning to work (P<0.001). At 1 year, symptom scores
were better in the surgical group (P=0.03). During the first year of
follow-up, there were 13 major adverse events in the embolization group
(12%) and 10 in the surgical group (20%) (P=0.22), mostly related to
the intervention. Ten patients in the embolization group (9%) required
repeated embolization or hysterectomy for inadequate symptom control.
After the first year of follow-up, 14 women in the embolization group
(13%) required hospitalization, 3 of them for major adverse events and
11 for reintervention for treatment failure.
Conclusions
In women with symptomatic fibroids, the faster recovery after
embolization must be weighed against the need for further treatment in
a minority of patients. (ISRCTN.org number, ISRCTN23023665.)
Uterine fibroids are the most common type of tumor in the female
reproductive system. The presence of these tumors may cause menstrual
disorder and can be associated with subfertility, miscarriage, and
pressure effects.1 For women who no longer plan to give birth, the
established treatment is hysterectomy. In the United Kingdom,
approximately 42,500 hysterectomies are performed annually, with
approximately 30% indicated for fibroids (the second-most-frequent
indication).2 For women wishing to maintain their fertility, myomectomy
is the principal option.
Uterine-artery embolization was introduced in 1995 as an alternative
technique for treating fibroids.3 Since then it has become increasingly
accepted as a minimally invasive, uterine-sparing procedure, and more
than 100,000 procedures have been performed during the past decade,
mainly in the United States and Western Europe.4 Early analysis of an
open, prospective, voluntary U.S. registry including 3160 patients
revealed major complications in 5.5% of patients at 30 days, with 0.1%
requiring a hysterectomy.5 In the United Kingdom, the National
Institute for Health and Clinical Excellence issued guidelines in
October 2004, stating that the procedure appeared to be safe for
routine use and that the majority of patients have short-term
symptomatic relief.6 However, there has been a need for a careful
assessment of the affects of the procedure on quality of life,
particularly in comparison with standard surgical approaches.7 We
designed a randomized trial comparing uterine-artery embolization and
surgery to assess quality of life and other outcomes at 1 year of
follow-up.
Methods
We conducted the trial in 27 hospitals in the United Kingdom. Each
hospital was associated with one of four regional centers. Patients
were randomly assigned to study groups from November 2000 through May
2004. The 12-month follow-up was completed in September 2005.
The study was approved by the Multicenter Research Ethics
Committee and local ethics committees at each center. All patients
provided written informed consent. Potential patients were provided
with written information describing the study and possible risks,
including the unknown effect of embolization on subsequent pregnancy.
Experienced interventional radiologists performed the embolization;
patients were referred to specialist centers from district units in
which embolization was not available. Hysterectomy and myomectomy were
performed at each local center.
Patients
Women at least 18 years old were eligible if they had one or more
fibroids of more than 2 cm in diameter that could be adequately
visualized with the use of magnetic resonance imaging (MRI), caused
symptoms (such as menorrhagia or pelvic pain and pressure), and were
considered by the patient’s physician to justify surgical treatment.
Exclusion criteria included a contraindication to MRI, severe allergy
to iodinated contrast media, subserosal pedunculated fibroids, recent
or ongoing pelvic inflammatory disease, pregnancy, and any
contraindication to surgery. There was no upper limit on the size or
number of fibroids.
Procedures
Patients were randomly assigned to study groups according to a
computer-generated schedule (permuted blocks) held by the trial
coordinator. Randomization was stratified by center and was performed
in a 2:1 ratio, with twice as many patients allocated to the
embolization group as to the surgical group. This design allowed better
characterization of the outcomes of the embolization procedure with
minimal reduction in statistical power. The methods of hysterectomy or
myomectomy was not specified; the choice between these options depended
on whether the patient wished to retain her uterus for fertility or
other reasons. Both operations were included, since virtually all
operations for fibroids are performed by the open route, allowing
appropriate comparison of out-comes. The technique for embolization was
also not specified, but both uterine arteries had to be embolized and
the particle size of the embolic agent was standardized (500 to 710 µm).
Outcome Measures
The primary outcome measure was quality of life, as assessed at
12 months on the Medical Outcomes Study 36-Item Short-Form General
Health Survey (SF-36), with scores ranging from 0 to 100, with higher
scores indicating better function. This assessment has been validated
in women with menorrhagia.8 Secondary outcomes included an assessment
of findings on the EuroQol-5D questionnaire, an instrument used to
measure preferences for certain health outcomes, including
hysterectomy,9,10 with a range of scores paralleling that of the SF-36;
an 11-point symptom score, ranging from -5 (markedly worse) to +5
(markedly better); the time until the resumption of usual activities; a
satisfaction score measuring whether patients would recommend the
procedure to a friend; a linear-analogue pain score at 24 hours; the
presence or absence of complications; and treatment failure, defined as
the need for subsequent intervention for symptom control, including
hysterectomy or repeated embolization.
Complications were graded with the use of the classification system
of the Society of Interventional Radiology, as recommended in the
Standards of Practice11 as follows: no therapy required or no
consequence (grade 1); nominal therapy required or no consequence,
including overnight admission for observation only (grade 2); therapy
required, including minor hospitalization of less than 48 hours (grade
3); major therapy required, including an unplanned increase in the
level of care or hospitalization for at least 48 hours (grade 4); and
permanent adverse sequelae (grade 5). Grades 1 and 2 were considered to
be minor; grades 3 through 5 were considered to be major. Two of the
investigators (a gynecologist and a radiologist) independently
categorized the grades of complications. In 56% of cases, the
investigators were in complete agreement; in 91% of cases, they were in
agreement to within one grade of complication. In discordant cases, the
worse grade was used. Major adverse events included any major
complication, a life-threatening event, initial or prolonged
hospitalization, and intervention required to preent permanent
impairment or damage, and death. Treatment failures requiring
subsequent intervention were considered separately.
We assessed outcome measures (with the exception of 24-hour pain
score) at 1, 6, 12, and 21 months and annually thereafter. In this
study, we present the 12-month results, with two exceptions: major
adverse events requiring hospitalization and subsequent intervention
for treatment failure, which are reported through September 2005
(maximum follow-up, 58 months).
Economic Analysis
We prospectively collected data on the total use of financial
resources up to 12 months after treatment. These data included the time
in the operating room, the total length of stay in the hospital,
outpatient visits associated with the procedure, treatment failure, and
any associated complications. We obtained unit costs for all resources
used from routinely collected data and published literature; we used
such data to determine the direct health care costs associated with
each patient from the perspective of the National Health Service. Since
the trial showed no significant differences between groups in the
primary outcome, we considered the appropriate form of economic
evaluation to be a cost-minimization analysis.12 We calculated the 95%
confidence intervals (CI’s) for the differences in cost between groups
with the use of the bias-corrected and accelerated bootstrap method.13
We performed one-way sensitivity analysis on key unit cost components
by varying on measure at a time.
Statistical Analysis
We analyzed all patients in the group to which they were randomly
assigned, regardless of the treatment actually received. Analysis of
covariance was used to compare quality-of-life scores (on the basis of
results on the SF-36 and EuroQol questionnaires) between groups,
adjusting for baseline values. Other comparisons between groups were
made with use of a two-sided Student’s t-test and the Mann-Whitney test
for continuous data and the chi-square test for categorical data. The
original power calculation required the enrollment of 200 patients to
give a power of 90% to detect a difference of 10 points in the SF-36
score at 12 months (the primary end point) at the 0.05 significance
level. Because of the slower-than-expected recruitment, the decision
was subsequently made to reduce the power to 80%, which required the
enrollment of 150 patients.
An independent data and safety monitoring committee reviewed the
results and serious adverse events every 12 months. The panel followed
the highly conservative Haybittle-Peto approach of requiring a
significance level of less than 0.001 in the comparison between groups
before making any recommendation to terminate the trial prematurely.14
The manufacturers of the embolic agents used in the study (William
Cook Europe, Cordis and the Biocompatibles) had no role in the design
of the study; data collection, analysis, and interpretation; or the
writing of the final report. The Writing Committee members assume
responsibility for the accuracy and completeness of the data and for
the overall content and integrity of the article.
Results
A total of 157 women were randomly assigned to study groups: 106 to
undergo uterine-artery Embolization and 51 to undergo surgery,
including 43 hysterectomies and 8 myomectomies (Fig.1). Eight patients
(5%) did not receive their allocated treatments (five in the
Embolization group and three in the surgery group). In addition, there
was one technical failure in the surgical group (a myomectomy converted
to a hysterectomy owing to technical difficulties) and there were three
technical failures in the embolization group (owing to difficulty in
the identification or catheterization of one or both uterine arteries).
All the hysterectomies and myomectomies were performed through an
abdominal incision. The groups were well matched at baseline (Table 1).
Primary Outcome
The primary outcome measure (the SF-36 quality-of-life score at 12
months) as available for 140 of the 157 women (89%). The results on
SF-36 and EuroQol at 1 and 12 months are shown in Table 2. There were
no significant differences between groups in any of the eight
components of the SF-36 at 12 months, although at 1 month, the
embolization group had significantly greater improvement in scores than
the surgery group for the physical function, social function, and
physical-role components.
Secondary Outcomes
Women in the surgical group had a significantly higher pain score at
24 hours (Table2). Symptom scores at 1 and 12 months after the
procedure were significantly better in the surgical group. At 12
months, the percentage of women who reported that they would recommend
their treatment to a friend was higher in both treatment groups (93% in
the surgical group and 88% in the embolization group) (P=0.32).
The median hospital stay after uterine-artery embolization was
significantly shorter than that after surgery (1day vs. 5 days,
P<0.001). The median time until patients could resume all recorded
usual activities was significantly lower in the embolization group
(Table 2).
Minor Complications
Minor complications were reported by 36 women (34%) in the
embolization group and 10 (20%) in the surgical group (P=0.06)
(Table3). Minor complications were usually related to the
postembolization syndrome (52%), which includes pyrexia, pain, and
elevated inflammatory markers, in the embolization group and to minor
infections (25%0 in the surgical group.
Major Adverse Events
There were 16 major adverse events (15%) in the embolization group,
as compared with 10 (20%) in the surgical group during a median
follow-up of 32 months (interquartile range, 23 to 41) (Table 3). When
we categorized these events with respect to the timing of their
occurrence (i.e., during the hospital stay, during the first year of
follow-up, or after the first year), 8 of the 10 major adverse events
in the surgical group occurred after discharge from the hospital.
Treatment Failures
Twenty-one patients (20%) in the embolization group required an
additional invasive procedure (hysterectomy or repeated uterine-artery
embolization) for continued or recurrent symptoms, 10 during the first
12 months of follow-up (2 of which were due to technical failures) and
11 subsequently. In the surgical group, there was one conversion of
myomectomy to hysterectomy at the time of the primary procedure.
Economic Analysis
Uterine-artery embolization was associated with the lower use of
resources then was surgery at the initial hospitalization. However,
during the 1-year follow-up period, when compared with surgery,
embolization was associated with more imaging studies and a longer mean
hospital stay.Table 4 shows the results of the cost-minimization
analysis and one-way sensitivity analysis.15-18 Uterine-artery
embolization was associated with total costs significantly lover than
those for surgery (mean difference, ₤951 [$1,712 at an exchange rate of
₤1 = $1.80]; 95% CI, ₤329 to ₤1480 [$592 to $2,664], suggesting that at
1 year, embolization was more cost-effective than surgery for patients
with symptomatic uterine fibroids, from the perspective of the National
Health Service. Sensitivity analyses showed the result was robust when
assumptions were varied around the cost of MRI and the embolization
agent. The results were sensitive to the cost per inpatient-day, with
no significant difference in costs between the two procedu
res
when the cost per inpatient-day was halved. Threshold analysis
indicated that uterine-artery embolization was more cost-effective over
a 12-month period only if the cost per inpatient-day exceeded ₤291
($524).
Other Outcomes
Through September 2005, eight pregnancies had occurred in five women
(seven in the embolization group and one in the myomectomy group). Four
of the pregnancies resulted in miscarriage, three in successful live
births (two by cesarean section, including one patient from each group,
and one spontaneous vertex delivery), and one intrauterine death of the
fetus at 33 weeks (with no abnormalities found on postmortem
examination).
Discussion
In this randomized trial comparing uterine-artery embolization with
standard surgical treatment for women with symptomatic fibroids, we
found no significant differences between the groups in measures of
quality of life at 12 months, although women in both groups had
substantial improvements in each component of the SF-36 score relative
to baseline. In contrast, the adverse-event profiles were very
different. Surgery was associated with the expected acute morbidity,
but only one major adverse event was recorded after the initial
hospital stay. Uterine-artery embolization was associated with a
significantly faster recover, including the resumption of usual
activities.
Rates of minor complications or major adverse events did not differ
significantly between the study groups, although the nature and timing
of these events varied between groups; major adverse events in the
surgical group typically occurred during the hospital stay, whereas in
the embolization group, such events more commonly occurred after
hospital discharge. Of note, three of the major adverse events in the
embolization group were cancers (two breast cancers, both detected
within 2 months after the intervention, and one adrenal cancer), which
were highly unlikely to be related to treatment.
At 1 year, however, 10 of the 106 women in the embolization group
had required a secondary intervention to treat persistent or recurrent
symptoms. After the first year of follow-up, 11 additional women were
readmitted for the same indication. These findings are consistent with
data from uncontrolled case series indicating complications and
treatment failures up to 48 months after embolization.
The cost-minimization analysis showed that at 1 year, embolization
was more cost-effective than surgery. This finding supports that of one
other study addressing the cost-effectiveness of uterine-artery
embolization versus surgery.21 Ongoing follow-up will further assess
the efficacy and cost-effectiveness of embolization.
We used a “pragmatic trial” design, in that the particular surgical
interventions and technical aspects of the procedures were not dictated
by protocol. We included women undergoing either hysterectomy or
myomectomy in the surgical group, although in fact only eight women
under-went myomectomy. Our primary outcome measure, the SF-36 score,
did not take specific fibroid-related symptoms into account, although
it was sensitive to changes in quality of life that resulted from
successful treatment of menstrual symptoms.8 This fact is important,
given the cyclical nature of the patients’ menstrual problems. We did
not collect data on loss of menstrual blood; comparisons of this
measure between groups would not be meaningful, given that only eight
women in the surgical group underwent myomectomy.
Two other randomized, controlled trials compared uterine-artery
embolization with hysterectomy.22,23 The first study used a
controversial randomized-consent methodology,24 in which women who were
randomly assigned to the hysterectomy group were not informed about the
study or the possibility of an alternative treatment (i.e.,
uterine-artery embolization). In addition, this study was small
(enrolling only 57 women) and used the length of hospital stay as the
primary outcome measure; hospital stays were significantly shorter
after uterine-artery embolization, with similar complication rates in
both groups.22 The second trial comparing embolization and hysterectomy
enrolled 177 patients; at 6 weeks after treatment, the embolization
group had a significantly shorter mean hospital stay but a higher rate
of minor complications and readmission.23
Limitations of our trial must be acknowledged. The original target
number of 200 patients was reduced to 150 because of difficulties in
recruitment. Thus, the 95% CIs for the differences between groups
indicate that plausible results include as much as a 10-point
difference between groups in some components of the SF-36. However,
there is no suggestion of clinically important differences. The
inclusion of only a small number of patients who underwent myomectomy
in the surgical group made it difficult to compare such therapy with
uterine-artery embolization. It also suggests that a direct comparison
of these two treatments would be difficult to perform unless
recruitment involved a very large population base. The use of the time
until resumption of usual activities as a secondary outcome must be
viewed cautiously, since such an interval could be biased by the
patient’s expectation (or caregivers’ guidance) regarding the time to
recovery.
The results of our study make clear that the choice between surgery
and uterine-artery embolization for symptomatic uterine fibroids
involves tradeoffs. The advantages of embolization — including a
significant reduction in the length of the hospital stay and 24-hour
pain level and a more rapid return to usual activities — need to be
weighed against the risk of treatment failure requiring a second
intervention and the possibility, although infrequent, of major late
adverse events. Longer-term follow-up is necessary, with attention to
the need for repeated intervention, to inform future decision making.