Description of included studies
Ten studies were included in this review (Colacurci et al., 1998; Dias Pereira et al., 1998; Fernandez et al., 1995; Fernandez et al., 1998; Hajenius et al., 1997; Moeller et al., 2009; Nieuwkerk et al., 1998; Saraj et al., 1998; Sowter et al., 2001b; Zilber et al., 1996).
All the included studies are randomised trials: three were conducted in the Netherlands (Hajenius et al., 1997; Dias Pereira et al., 1998; Nieuwkerk et al., 1998), two in France (Fernandez et al., 1995; Fernandez et al., 1998), one in Denmark (Moeller et al., 2009), one in the USA (Saraj et al., 1998;), one in Israel (Zilber et al., 1996), one in New Zealand (Sowter et al., 2001b) and one in Italy (Colacurci et al., 1998).
All included studies compared medical and surgical management of ectopic pregnancy and reported at least one priority outcome. Surgical management consisted of salpingotomy/salpingostomy in all of the included trials except for one (Sowter et al., 2001b) where some women (n = 7/28) were treated with salpingectomy for various reasons. Medical management in all included trials consisted of systemic or local injection of methotrexate. Eight trials (Colacurci et al., 1998; Dias Pereira et al., 1998; Fernandez et al., 1998; Hajenius et al., 1997; Moeller et al., 2009; Nieuwkerk et al., 1998; Saraj et al., 1998; Sowter et al., 2001b) compared systemic methotrexate with surgery. In three of these trials (Nieuwkerk et al., 1998; Dias Pereira et al., 1998; Hajenius et al., 1997) women were treated with multiple doses of methotrexate and in the remaining five trials they were treated with a single dose of intramuscular methotrexate injection. Three trials (Fernandez et al., 1995; Fernandez et al., 1998; Zilber et al., 1996) compared local injection of methotrexate to surgery for management of ectopic pregnancy. Local injection in the three trials consisted of direct injection of methotrexate into the pregnancy site under laparoscopic guidance. One trial (Fernandez et al., 1998) compared local or systemic injection of methotrexate to surgery for management of ectopic pregnancy.
The trials were conducted in economically developed countries and their populations include women with confirmed ectopic pregnancy. Further details, including the duration of follow-up, can be found in the evidence tables in Appendix H.
Evidence profile
Outcomes are reported in three evidence profiles:
GRADE summary of findings for comparison of surgery with systemic methotrexate.
GRADE summary of findings for comparison of surgery with local methotrexate.
GRADE summary of findings for comparison of surgery with systemic and local methotrexate.
In the outcomes with high heterogeneity (I2 greater than 60%), the technical team used a random effects model (the remaining outcomes used fixed effects models) and explored the heterogeneity with sensitivity analyses.
Success rate (surgery compared with systemic methotrexate)
Overall heterogeneity (69%) was not reduced by excluding the trial with multi-dose methotrexate; however, the heterogeneity was reduced to some extent (56%) by excluding the trial that treated some women (n = 6/28) with salpingectomy.
Hospital stay (surgery compared with local methotrexate)
Excluding the study with small sample size (less than 50) reduced the heterogeneity from 100% to 0%. The change could be a result of different hospitals' policies regarding the hospital stay following the treatment.
Health economics
A new economic model was developed for this guideline to assess the cost effectiveness of different treatment strategies for ectopic pregnancy. The model is described in more detail in Section 10.4.
The model compared the cost effectiveness of three treatment strategies in a population of women diagnosed with an ectopic pregnancy but not requiring urgent surgical intervention:
In the base–case analysis the evaluation took the form of a cost minimisation analysis with the assumption that all women recover and that any differences in morbidity only exist in the very short term. Methotrexate was the cheapest option at £1432 followed by laparoscopic salpingectomy at £1608. Laparoscopic salpingotomy, because of its relatively high re-intervention rate and follow-up costs, was the most expensive strategy at £2205. A probabilistic sensitivity analysis of one million Monte Carlo simulations found that methotrexate was cheapest in 99.65% of the simulations with a 0.35% probability that laparoscopic salpingectomy was cheapest.
A number of one-way sensitivity analyses suggested that the ordinal ranking of strategies in terms of their cost was not affected by large changes in parameter values. A further sensitivity analysis relaxed the assumption about equivalence in treatment outcomes and estimated an incremental gain in quality adjusted life years (QALYs) for the surgical alternatives when compared to methotrexate. This found that laparoscopic salpingotomy was dominated by laparoscopic salpingectomy (a higher cost than laparoscopic salpingectomy with no QALY gain). The incremental cost effectiveness ratio of laparoscopic salpingectomy against methotrexate was calculated to be almost £84,000 per QALY which would not normally be considered cost effective using advisory willingness to pay thresholds suggested by NICE (NICE, 2009).
Evidence to recommendations
Relative value placed on the outcomes considered
The guideline development group (GDG) identified future reproductive outcomes as very important from the woman's perspective.
The group also felt that outcomes such as the chances of treatment success, the likelihood of needing another intervention and the length of time required for follow-up would be important considerations for women, as well as having cost implications.
Women's experiences of care and psychological outcomes were identified as being important. However, evidence was only found for psychological outcomes and was not available for women's experience of care. The effects demonstrated were short term, with no significant differences shown after 2 weeks in most cases and 4 months in the remainder.
Consideration of clinical benefits and harms
The evidence in this review showed that women undergoing medical treatment required a longer period of recovery and follow-up than women who had laparoscopic surgery. In some women, the time to resolution was up to 6–7 weeks. Medical treatment was also associated with a higher need for further intervention, which could include surgery or an additional dose of methotrexate. In addition, the GDG recognised that some women may have contraindications to surgery or have a personal desire to avoid a surgical procedure; therefore medical treatment would be appropriate in such circumstances. For women choosing to have methotrexate treatment, the GDG agreed that systemic administration was preferable, as it is easier to administer, less invasive and in line with current practice.
The GDG noted that the studies considered in this review had specific inclusion criteria for women with an ectopic pregnancy who were felt to be at low risk for tubal rupture. Thus they felt that for women who fell outside of these criteria, methotrexate treatment was not an appropriate choice of treatment. Evidence from Sowter et al. (2001b) showed that the chance of treatment failure was increased in women with higher initial hCG levels. At an initial hCG below 1000 international units per litre (IU/l), women had a 12% chance of requiring further treatment, but at an hCG over 1500 IU/l this rose to 70%. Given this finding, the group recommended that methotrexate only be offered as a first-line treatment for women with an hCG of less than 1500 IU/l (and with an ectopic pregnancy with an adnexal mass smaller than 35 mm).
The evidence from Sowter et al. (2001b) also showed that at an hCG level below 5000 IU/l, women had an 85% chance of successful treatment with either a first or second dose of methotrexate administration. Given this, they felt that for women with hCG levels between 1500 IU/l and 5000 IU/l, it would still be appropriate to offer medical management, but that surgery should also be offered as a choice, given that women in this group were more likely to require further interventions following medical management.
The GDG members agreed that for women with an hCG level greater than 5000 IU/l, with an adnexal mass greater than 35 mm or with an ectopic pregnancy with a fetal heartbeat visible on ultrasound, surgery should be offered as a first-line treatment, as they were aware of studies that showed an increased risk of rupture for all of these women.
The GDG was aware of cases of women who had experienced a failure of medical treatment and then re-presented later with tubal rupture or other severe symptoms: it was felt, therefore, that follow-up should be mandatory after any medical treatment and that women needed access to 24 hour emergency care. Using their clinical experience, the GDG members agreed that women should have two follow-up appointments in the first week following treatment, and then one appointment per week until a negative pregnancy test is performed. This follow-up protocol would ensure that women in whom medical treatment was not effective would be identified early, before their condition became serious. The GDG agreed that an hCG level which plateaued or rose following medical management should prompt a reassessment of the woman's treatment. Given the potential risks if medical management fails, the group highlighted the importance of this follow-up protocol, and decided that if any difficulties in follow-up were anticipated, women should be advised to have surgery as a first-line treatment.
Consideration of health benefits and resource uses
The GDG noted that, while surgery has the advantage of a shorter resolution time and a reduced need for further intervention, medical management is associated with a shorter stay in hospital. Therefore, the group decided that this area was appropriate for a health economics analysis. Using a cost minimisation approach, the model determined that methotrexate was the preferred treatment approach. Taking this finding into account, but also the potential risks inherent in tubal rupture, the GDG decided to recommend methotrexate as the first-line treatment for women meeting the criteria discussed above; namely those with an unruptured ectopic pregnancy which is smaller than 35 mm with no visible heartbeat and who have an hCG level less than 1500 IU/l.
Quality of evidence
The evidence was derived from randomised controlled trials and ranged in quality from high to low for the outcomes considered. The GDG felt that it was disappointing that studies had not reported more information about outcomes among women with different initial hCG levels and size of ectopic pregnancy. Four studies only included ectopic pregnancies below a certain size (ranging from 35 mm to 40 mm) and only two discussed the effect of hCG values on outcomes.
Information giving and emotional support
The GDG emphasised the importance of providing good quality information to women who are in a clinically stable condition, including accurate information on local management options and appropriate counselling regarding the proposed procedure. The GDG was aware that, for some women, methotrexate might be a clinically appropriate treatment option but would not be acceptable to the woman personally. In these circumstances, the GDG felt that the woman's choice should be supported. For all women with an ectopic pregnancy, the GDG thought that it was important that women were given sufficient information about what to expect during the course of their treatment and recovery. In particular, GDG members felt women should be made aware of the amount of pain and/or bleeding that might be expected, so that women knew that it was a normal part of the treatment process.
It was recognised that ongoing psychological support should be offered from the diagnostic period and during discussions regarding treatment options and the relevant outcomes and into the postoperative period, including information regarding resuming usual activity and details of patient groups to contact for support. The information needs to cover details regarding both physical and emotional recovery.
Other considerations
The GDG noted that these studies looked specifically at women with unruptured ectopic pregnancies and therefore these women were not in extreme pain. The GDG did not feel that women in significant pain were suitable for medical management and therefore they should be offered surgery.
The group also wished to highlight that medical management should only be undertaken once it was confirmed on an ultrasound scan that there was no intrauterine pregnancy.
