- Top of page
Objective To evaluate the role of pretreatment with gonadotrophin releasing hormone (GnRH) analogues (GnRHa) prior to surgery for women with uterine fibroids.
Design A systematic review was undertaken of all randomised controlled trials assessing the efficacy of GnRHa treatment prior to surgery (myomectomy or hysterectomy) in women with fibroids.
Sample Premenopausal women with fibroids awaiting surgery attending hospitals or clinics.
Methods Electronic searches of Medline, EmBase, Current Contents, the Cochrane Library, National Research Register, NLM's Clinical Trials Register and the Cochrane group's Trials Register were performed, between 1980 and 2000. Twenty-six randomised controlled trials were identified that compared GnRHa with no treatment, placebo or other medical treatment. The quality of the trials was assessed and data were extracted independently by two of the reviewers. Statistical analysis was performed in Revman according to Cochrane guidelines and where possible outcomes were pooled in a meta-analysis.
Main outcome measures Pre-operative assessment (reduction in uterine and fibroid volume, change in haemoglobin and haematocrit, change in patient's symptoms), intra-operative assessment (duration of operation, blood loss, proportion with vertical incision, proportion undergoing vaginal rather than abdominal procedure, frequency of blood transfusions), post-operative assessment (complications, duration of hospital stay, recurrence of fibroids).
Results Pre- and post-operative haemoglobin and haematocrit were significantly improved by GnRHa therapy prior to surgery, and uterine volume, uterine gestational size and fibroid volume were all reduced. Pelvic symptoms were also reduced but some adverse events were more likely during GnRHa therapy. Hysterectomy appeared to be easier after pretreatment with GnRHa therapy; there was reduced operating time and a greater proportion of hysterectomy patients was able to have a vaginal rather than an abdominal procedure. Duration of hospital stay was also reduced. Blood loss and rate of vertical incisions were reduced for both myomectomy and hysterectomy. Evidence of increased risk of fibroid recurrence after GnRHa pretreatment in myomectomy patients was equivocal and few data were available to assess change in post-operative fertility.
Conclusions The use of GnRHa for three to four months prior to fibroid surgery reduces both uterine volume and fibroid size. They are beneficial in the correction of pre-operative iron deficiency anaemia, if present, and reduce intra-operative blood loss. If uterine size is such that a midline incision is planned, this can be avoided in many women with the use of GnRHa. For women undergoing hysterectomy, a vaginal procedure is more likely following the use of these agents.
- Top of page
Uterine fibroids are benign tumours of the smooth muscle cells of the uterus. They are found in at least 20–25% of women over the age of 35 years1 although they may be as frequent as 50%2. It is estimated that 20–50% of these tumours cause symptoms, mainly excessive menstrual loss, infertility, recurrent pregnancy loss and pelvic pressure, severe enough to warrant therapy3,4. The standard treatment for symptomatic uterine fibroids has always been surgical, either hysterectomy or, in women who wish to preserve their fertility, the more conservative procedure of myomectomy. Fibroids represent one of the most frequent indications for major surgery in premenopausal women5, and as such, they constitute a major public health cost.
It has been recognised that fibroid growth and maintenance are stimulated by oestrogen and are affected by hormonal cyclic changes6. Oestradiol and progesterone receptors have been identified in myomatous tissue7–9 and thus fibroids may be responsive to therapeutic hormonal manipulation.
Since the 1980s, gonadotrophin releasing hormone (GnRH) agonists, which induce a state of hypo-oestrogenism by suppressing pituitary ovarian function, have been used as a treatment for fibroids. The main effects of this treatment are the temporary control of bleeding and a reduction in fibroid and uterine size. However, side effects include menopausal symptoms and bone loss with long term use. After therapy is stopped, there is regrowth of both the fibroids and the uterus almost to their pretreatment size and a recurrence of symptoms in a majority of patients10.
Although the use of these agents as a primary medical therapy for fibroids has been disappointing, the shrinkage of fibroids and the uterus during treatment with GnRH analogues (GnRHa) and the reduction in uterine blood flow have led to investigation of their role as a pre-operative adjunct to surgery.
Pretreatment of patients with GnRHa prior to hysterectomy has been advocated for patients with severe anaemia and in order to reduce blood loss during surgery. Other indications have included large fibroids or other factors that would make surgery technically difficult11. Pretreatment may also enable the greater use of vaginal rather than abdominal hysterectomy12 or even more conservative surgical options such as laparoscopic hysterectomy.
Myomectomy has generally been used for women with a desire to preserve or enhance their fertility but is often regarded as a more difficult procedure than hysterectomy, with a higher risk of post-operative complications. Intra-operative haemorrhage can also necessitate blood transfusion or hysterectomy. Myomectomy may be performed via laparotomy, laparoscopy or hysteroscopy and the method must be distinguished in the evaluation of pretreatment with GnRHa. Potential benefits are the reduction in blood loss during the operation, ease of operability, better anatomic reconstruction and the possibility of utilising a transverse (Pfannenstiel-type) rather than vertical midline incision at laparotomy. However, concern has been expressed that the fibroid capsule will become less evident and may be missed, fibroids will not ‘shell out’ cleanly and the tumour excision may be more difficult13,14.
The aim of this systematic review is to evaluate the role of pretreatment with GnRHa prior to a major surgical procedure, either hysterectomy or myomectomy, for uterine fibroids.
- Top of page
All publications which described randomised trials of GnRHa administered prior to surgery for the treatment of uterine fibroids were obtained using the search strategy developed by the Cochrane Menstrual Disorders and Subfertility Group for the Trials Register. The Trials Register contains both published and unpublished trials identified from regular searches of Medline, EmBase, Cochrane Library, and handsearching specialist journals and conference abstracts. Current Contents, National Research Register, the National Library of Medicine's Clinical Trials Register and reference lists of identified trials and relevant review articles were also searched. Drug companies were contacted for details of unpublished trials.
The selection of trials for inclusion in the review was performed by two of the reviewers. Included trials were analysed for quality criteria that included method of randomisation, blinding to treatment allocation, quality of allocation concealment, number of withdrawals and whether an intention-to-treat analysis and power calculation for sample size were performed. All assessments of the quality of trials and data extraction were performed independently by two reviewers using forms designed according to Cochrane guidelines. Two of the reviewers were experts on clinical issues and one reviewer had statistical expertise. Where necessary, additional information on trial methodology or actual original trial data were sought from the corresponding author of any trials that appeared to meet the eligibility criteria.
There are considerable differences between the types of surgical procedure used in the treatment of uterine fibroids in relation to subsequent outcomes from GnRHa pretreatment. Because pre-operative outcomes such as uterine volume are not affected by the type of surgery subsequently performed, all the results of pre-operative outcomes were combined for both major types of surgery, hysterectomy and myomectomy in the review. Intra- and post-operative outcomes, however, were assessed separately according to whether myomectomy or hysterectomy was performed. Subgroup analysis was not performed according to the route of surgery for each type, for example, abdominal, vaginal or laparoscopic route for hysterectomy, as data were not available. Subgroup analysis, however, was performed according to uterine size in gestational weeks. Data were available separately for two subgroups in one study, uterus 14 to 18 gestational weeks and greater than 18 gestational weeks, although for most of the included studies, uterine size was either not an inclusion criteria or specified as greater than 12 weeks of gestation. In addition, subgroup analysis was performed according to type of control group comparison, whether placebo controlled or immediate surgery with no pretreatment. Where results from these two types of control groups differed, the possibility of bias was considered. Some outcomes such as difficulty of surgery and withdrawal because of adverse effects were more likely to be subject to bias from lack of blinding. Other outcomes such as haemoglobin values are unlikely to have been affected by whether the control group was placebo controlled or not. No other subgroup analysis was performed but sensitivity analysis was performed with trials where groups were not comparable at baseline, with trials where dosage regimens and route of administration differed, with trials that had skewed data and with trials where pre-operative iron supplementation was included in the intervention.
Statistical analysis was performed in accordance with the guidelines for statistical analysis developed by the Cochrane Menstrual Disorders and Subfertility Group15. Where possible, the outcomes were pooled statistically. Heterogeneity between the results of different studies was examined by inspecting the scatter in the data points and the overlap in their confidence intervals and, more formally, by checking the results of the χ2 tests. A priori, it was planned to look at the possible contribution of differences in trial design to any heterogeneity identified in this manner.
For dichotomous data (for example, proportion of patients with post-operative complications), results for each study were expressed as an odds ratio with 95% confidence intervals and combined for meta-analysis with RevMan software using the Peto-modified Mantel–Haenszel method.
For continuous outcomes, a fixed effects model was used and where data were pooled a weighted mean difference with 95% confidence intervals was reported. For many studies, the outcome data, particularly intra-operative blood loss or duration of surgery, were skewed and authors correctly presented the summary effects as medians with a range. Some studies reported a mean change or percentage change from baseline values. Where possible, original data were obtained from the principal authors, but post-treatment means and standard deviations were not always either available or calculable. Where only medians and ranges were available, the median was regarded as being identical to the mean and an estimate of the standard deviation was calculated from the range [(range × 0.95)/4]. Sensitivity analysis was performed with and without trials with skewed data included in the meta-analysis and results were compared.
Twenty-six randomised controlled trials of presurgical treatment with GnRHa in women with fibroids were identified. Studies of GnRHa treatment in women with fibroids where surgery was not reported were not considered in this review. One randomised controlled trial of GnRHa pretreatment compared with other medical pretreatment (lynestrenol) was identified. Five trials meeting the inclusion criteria are still awaiting assessment and their results are not included in the review. The data were not provided separately for patients with fibroids in one study [Ylikorkala, 1995], extra information was required for two studies [Benagiano, 1992; Reinsch, 1994] and two trials are awaiting translation [Gambardella, 1995; Nakano, 1998].
Fourteen randomised controlled trials of GnRHa therapy vs no therapy with a total of 1005 patients, six randomised controlled trial of GnRHa therapy vs placebo with a total of 825 patients and one randomised controlled trial of pre-operative GnRHa therapy vs lynestrenol met the criteria for inclusion in the review (Table 1). In studies with a no treatment arm, hysterectomy was performed in four of the trials, myomectomy in five trials, a combination of myomectomy and hysterectomy in four trials and unspecified surgery in one trial [Cagnacci, 1994]. In studies with a placebo arm, hysterectomy was performed in four trials, myomectomy in one trial and a combination of hysterectomy and myomectomy in one trial. The other trial, comparing GnRHa with another medical pretreatment, lynestrenol, used a combination of myomectomy and hysterectomy [Verspych, 2000].
All patients were premenopausal and had symptomatic fibroids (i.e. the presence of fibroids was reported as confirmed by ultrasonography in 20 of the studies). Ovarian pathology was an exclusion criterion in 12 studies. In the two placebo studies with iron in the treatment arms, patients were only included if they had iron deficiency anaemia.
A number of different GnRHa preparations were administered via different routes and different regimens. Leuprolide acetate, goserelin and triptorelin were given either by intramuscular depot injection or subcutaneous depot implant every four weeks prior to surgery and in three studies nafarelin or buserelin were given daily by nasal spray.
Duration of treatment ranged from two to three months and in two studies patients were treated for four months. In those trials with a no pre-operative treatment arm, the patients in the control group had surgery either immediately or as soon as practicable, but in two studies [Cagnacci, 1994; Nikolov, 1997] there was a three-month observation period prior to surgery equivalent to the duration of treatment in the GnRHa group.
Dosage for the depot formulations was either 3.6 mg (goserelin), 3.75 mg (leuprolide acetate or triptorelin) or 3.2 mg (triptorelin) but in one study [Stovall, 1995], comparisons were made between two different doses of leuprolide acetate, 3.75 mg vs 7.5 mg. Data from the higher dose were not initially entered in the review as all other doses were equivalent but sensitivity analysis enabled the comparison of results with and without the inclusion of the larger dose. Two of the placebo trials included iron in both treatment arms as patients were anaemic [Stovall, 1995; Benagiano, 1996]; the latter trial also included a GnRHa + placebo iron arm, which was not considered in this review. Sensitivity analysis was also performed with and without the inclusion of the studies with iron treatment and the studies with intranasal GnRHa administration in the meta-analysis to determine if results varied.
Pre-operative uterine and/or fibroid volume was calculated either by the prolate ellipsoid method and the formula V= 0.5233 (D1 × D2 × D3), where D1, D2 and D3 are the longitudinal, transverse and cross sectional diameters of the uterus or fibroid, respectively16 or the water displacement method in 17 studies. Uterine size in gestational weeks was assessed in four studies. Intra-operative blood loss, reported in 17 studies, was estimated by measuring the weight of swabs and the volume of blood collected into receptacles such as aspiration bottles. In two placebo-controlled studies and one comparative study, the proportion of patients with pelvic pain or pressure was recorded and in two studies with a no treatment arm and one placebo study this outcome was recorded in a score as either none, mild, moderate or severe on a scale of 0 to 3. Women recorded the presence/absence and severity of their myoma-related symptoms prior to treatment and the improvement in these symptoms after treatment in one study [Friedman, 1992]. Another study [Shaw, 1996] assessed improvement in dyspareunia and dysmenorrhoea after pre-operative treatment.
Other outcomes included pre-operative haemoglobin and haematocrit (15 studies), post-operative haemoglobin and haematocrit (seven studies), duration of operation (15 studies), duration of hospital stay (nine studies), frequency of blood transfusions (13 studies), type of incision (five studies), degree of difficulty of the surgery (six studies), post-operative complications (11 studies), frequency of adverse events (nine studies), proportion undergoing vaginal hysterectomy (three studies), treatment acceptability (four studies), change in fertility status (one study) and recurrence of myomas (two studies). No data were provided for the outcome of intra-operative hysterectomy (for women undergoing myomectomy) in the included studies.
Four of the placebo trials, three of the no treatment trials and the study comparing GnRHa with lynestrenol had adequate concealment prior to allocation. Concealment of allocation was not adequately reported in the remainder of the randomised trials and one trial used alternation to allocate women to groups. There were marked differences in baseline parameters between groups in six studies that were likely to affect the assessment of outcomes as a result of treatment.
Five of the six placebo trials used double blinding (it is likely that it was also used in the other trial but this was not specifically reported). Although none of the patients were blinded in the trials with a no treatment arm before surgery, the surgeon was blinded in three and the ultrasonographer was blinded in one study. Power calculations for sample size were performed for four placebo and two no treatment studies and analysis by intention-to-treat in four placebo, three no treatment studies and the comparative trial using lynestrenol. One trial was preceded by a pilot study [Lumsden, 1994].
Subgroup analysis was provided separately for patients with uterine size <600 cm3 and size ≥600 cm3 in one study [Friedman, 1989] but numbers were very small. Another study assessed duration of operation in three separate subgroups, number of fibroids greater than and less than three, volume of main fibroid greater than and less than 60 mL and echogenicity score (scored from 1 to 7) of main fibroid greater than and less than three [Zullo, 1996]. Data from all of these separate analyses have not been included in this review. Data from another trial [Campo, 1999], which followed up a subgroup of infertile participants for 13 months after surgery to assess the rate of spontaneous pregnancies, were included in the analysis as this was a prespecified outcome in the trial and this review. However, numbers in each group were small and there was no evidence of methods used to control for confounding. A placebo-controlled trial compared different doses of leuprolide acetate (3.75 vs 7.5 mg) in different strata (baseline haematocrit ≤28% and >28%) and sensitivity analysis was performed to assess the effects of different GnRHa doses and different strata on results. One trial of anaemic patients compared two different treatment groups (goserelin + placebo iron and goserelin + daily iron) with a placebo + daily iron arm. Data from the first treatment group were not considered in this review as differences in results could be due to the lack of iron treatment. Data from the intra- and post-operative outcomes in two trials [Golan, 1993; Shaw, 1989] were reported separately according to the type of surgery performed, myomectomy or hysterectomy. Subgroup analysis (in patients with uterine size between 14 and 18 gestational weeks and greater than 18 gestational weeks) was also performed in another trial with a no pre-operative treatment arm [Stovall, 1994] and was considered separately in the meta-analysis.
Ten of the no treatment studies and one placebo study had no reported withdrawals after randomisation. Of the remaining placebo trials, withdrawals prior to treatment ranged from 1% to 14%, two studies had 5% and 15% withdrawals during treatment, one study had 7% withdrawal prior to treatment and a further 7% prior to completion and one study had 3% withdrawal by the end of the study. One placebo trial had a total withdrawal prior to surgery of 29% [Stovall, 1995]. Three of the no treatment studies had withdrawals that ranged from 3% to 9% after randomisation and before GnRHa treatment. Two studies had 13% withdrawal before surgery and one study had 13% withdrawal after surgery. This latter study reported a total withdrawal of 34% after randomisation. Another trial had 18% withdrawal prior to completion of pre-operative treatment. Reasons for withdrawals were usually due to unwillingness to continue, failure to meet some inclusion criteria, adverse events (in the treatment group), personal reasons and pregnancy or menopausal symptoms. Sensitivity analysis was performed with and without the inclusion of the two studies that had a withdrawal rate greater than 20% to check if results varied.
- Top of page
This review has assessed the effect of GnRHa on a number of important pre-operative, peri-operative and post-operative outcomes for both myomectomy and hysterectomy in the surgical treatment of uterine fibroids. It has assessed both the likely benefits and harms of this treatment, but has not addressed the issue of whether it is cost effective.
The combined results of both placebo-controlled and GnRHa vs no treatment trials strongly suggest that the use of GnRHa is associated with an increase in pre-operative haemoglobin and haematocrit. This may be important in women with severe menorrhagia and anaemia, although a weighted mean increase of 1.3 g/dL for haemoglobin concentration and 3.1% for haematocrit may be of relatively little clinical importance in many cases.
Both uterine and fibroid volume, however measured, are very significantly reduced by the use of GnRHa. The weighted mean difference when data from these studies are combined represents an average reduction in uterine volume of 159 mL and this reduction may be greater in women with a larger uterus. This finding is supported by a significant reduction (two weeks in gestation) in uterine gestational size. Fibroids were reduced in size by an average of 12.5 mL. GnRHa pretreatment was also more successful than lynestrenol in reducing fibroid volume. The effect on fibroid volume was not stratified for pretreatment size in any trial. It is possible that a greater relative reduction is seen in women with larger fibroids and this should be investigated.
There is consistent evidence that GnRHa are associated with an improvement in pelvic pain and symptom scores. These beneficial changes must be weighed against the significant proportion of women who experience adverse events, related to hypo-oestrogenism, such as vaginitis and hot flushes, which can affect adherence to therapy.
One of the most important areas of uncertainty that these studies have attempted to clarify is whether the use of GnRHa actually makes surgery easier. Parameters such as operating time, total blood loss and post-operative haemoglobin concentration provide an objective, although indirect, measure of operative difficulty. Parameters such as operative difficulty as rated by the operating surgeon, and type of incision used may be prone to bias when treatment allocation is known.
Combining data from trials with a no treatment arm shows that intra-operative blood loss is reduced at both hysterectomy and myomectomy. In the one small placebo-controlled trial reporting this, a non-significant trend favouring GnRHa is seen and data from larger double-blind placebo-controlled trials would be useful. These reductions in blood loss are associated with a small improvement in post-operative haematologic indices, although no benefit in terms of a reduced transfusion requirement has been demonstrated (not unexpected, given the relatively small number of women included into trials that have evaluated these parameters).
The few trials that have reported operating time (a reasonable indirect measure of operative difficulty) for myomectomy surgery have shown no benefit associated with their use. For hysterectomy a significant reduction in operating time is shown for those trials comparing treatment with no treatment and a significant trend in favour of GnRHa for placebo trials. However, the reduction in operating time is likely to be only about 5 minutes or less than 10% of the total operating time.
In the studies where it has been reported, both the choice of surgical approach (vaginal vs abdominal) and type of incision (midline vs transverse) are altered favourably with the use of GnRHa, presumably as a consequence of the reduction in uterine volume. This suggests that these agents may be useful pre-operatively for women where an abdominal hysterectomy or midline incision is being proposed because of uterine size.
Five studies have rated operative difficulty during hysterectomy. The placebo-controlled trials show a significant reduction in the proportion of GnRHa-treated patients with difficult surgery but overall any beneficial effect on operative difficulty appears to be small. No data are available for myomectomy, and the issue of whether GnRHa make the process of ‘shelling out’ a fibroid difficult has not been answered by these trials.
Post-operative complication rates for both hysterectomy and myomectomy are unaffected by the use of GnRHa, although there may be a small reduction in the duration of post-operative stay amounting on average to less than a day.
One important issue associated with the use of GnRHa for myomectomy is whether they result in an increased risk of fibroid recurrence, presumably because small fibroids are not seen at the time of surgery. Two studies have evaluated this; one small unblinded trial showing an adverse affect due to GnRHa and one small placebo-controlled trial showing no difference in recurrence risk associated with their use. At present, this question cannot yet be adequately answered by randomised data.
Another as yet unanswered question is whether GnRHa improve fertility outcomes after myomectomy; this question has been addressed by only one small randomised trial. Any differences in cumulative pregnancy rates after myomectomy are likely to be small but follow up data on fertility outcomes for women entered into these and future trials would be useful.
Because there are clearly established advantages and disadvantages of GnRHa, it is useful to compare it with other types of pretreatment. One advantage of fibroid or uterine reduction is the possibility it offers women to gain access to other types of surgery and benefits such as shorter surgery, less blood loss and shorter recovery time. GnRHa treatment in one small randomised controlled trial was more successful at reducing fibroid volume and resulted in a lower reduction in blood parameters such as haematocrit post-operatively than a progestogen, lynestrenol. Women were more likely to suffer hot flushes and headaches when pretreated with GnRHa but fibroid-related symptoms showed similar improvement in both groups. Thus, lynestrenol did not appear to be as effective as GnRHa as pre-operative treatment before surgery for fibroids.
Implications for practice
There is clear evidence from randomised trials that the use of GnRHa is associated with a significant reduction in uterine volume and fibroid size. Haematologic indices are improved slightly and there may be some reduction in intra-operative blood loss and operating time. However, there is inadequate evidence to support the use of GnRH agonists for all women with fibroids undergoing hysterectomy or myomectomy. They can be recommended for pre-operative use in women with a greatly enlarged uterus, pre-operative anaemia or where a midline rather than transverse incision would be planned. In addition, some women undergoing hysterectomy will be able to benefit from a less invasive vaginal rather than an abdominal procedure.
Implications for research
To date, few blinded placebo-controlled data are available evaluating operative difficulty and future trials should be both blinded and should consider evaluating operative difficulty in a reproducible way.
Cost effectiveness data are lacking in all trials published so far and, given the very significant cost of these agents, some attempt should be made to include such data in future trials.
The question of whether the chances of fibroid recurrence is increased after the use of GnRHa and whether subsequent fertility outcomes are improved by their use should be evaluated further within future randomised trials.