Introduction

After fertilisation of the oocyte with a spermatozoon, the blastocyste implants around the sixth/seventh day in the endometrial layer of the uterus. In 2 % of the cases, the nidation takes place outside the uterus (extrauterine gravidity = EUG). When this occurs 96 % of the cases are found in the tube. Other localisations can be in the ovaries, cervix, abdominal cavity or uterus walls (i.e. Caesarean section scars [1]). With the technological improvement of ultrasound and structured teaching programs of the colleges (i.e. Deutsche Gesellschaft für Ultraschall in der Medizin (DEGUM), recommendations of the deutsche Gesellschaft für Gynäkologie und Geburtshilfe (DGGG) for ultrasound [2]) in conjunction with human chorionic gonadotropin (HCG) level diagnosis is often possible on asymptomatic patients [3].

Depending on the clinical findings, therapy of the EUG can be done either medically (commonly with methotrexate) or surgically with the removal of the pregnancy tissue [4]. A Cochrane review [5] concludes for the small ectopic pregnancy a similar outcome. The long term success rate did not differ significantly in the follow-up. The long term outcome was defined as further conceptions and pregnancies. An additional result was the equivalence of the different medical protocols and dosages [5]. The authors conclude that in the daily routine a clinically stable patient with low HCG levels and a sonographically detected EUG, the medical option should be preferred.

Due to increasing cost-effective concerns of health insurances, care providers and patients, the optimal treatment is not known from an overall perspective. This article systematically reviews the current literature in regard to costs of the laparoscopic and medical treatments.

Methods

Following the PRISMA guidelines, the National Library of Medicine online database was searched (pubmed) using ((extrauterine or (extra uterine) or ectopic) and pregnancy and treatment and cost). This resulted in 260 articles. Language was limited to English and publications before 1.1.1997 were excluded (Fig. 1).

Fig. 1
figure 1

Literature selection with exclusion criterias

Full size image

Amongst the 17 results were two systematic reviews from Hajenius und Mol [5, 6]. Both publications reviewed randomised controlled trials (RCT) from January 1966 to February/October 2006/September 2007. The main focus was on the follow-up pregnancies. The articles quoted for the cost analysis were also returned by the performed pubmed search.

Two further publications [7, 8] were comments to an older article [9], three articles [10–12] compared the cost of laparotomy with the costs of laparoscopy. A further two articles [13, 14] simulated a cost analysis with previous published data. Barnhart [15] analysed the cost for the different medical regimes and an observational study from Seror [16] republished data from the same database as Vaissade [17].

Therefore six clinical trials remained (listed in Tables 1, 2) for the cost analysis of extrauterine gravidity treatment listing the direct and partly indirect costs.

Table 1 Details of the included studies
Full size table
Table 2 Study details and results
Full size table

Results

Since 1.1.1997, six relevant publications were identified, publishing information on costs of the minimal invasive and medical treatment costs. Three of these were retrospective studies from epidemiological databases [18–20], the other three were prospective trials [16, 21, 22]. Details are given in Table 1.

The study populations varied depending on the primary questions regarding diagnosis, medical regime used and follow-up. For example in the database analysis [16], acute/sub-acute EUGs were differentiated and the therapy was listed accordingly. The studies [21, 22] looked at all cases with the diagnosis of EUG and excluded pregnancies with unclear location, HCG levels >5,000 IU/L, sonographic diameter >3.5 cm or patients with unstable vital parameters.

Inclusion criteria for all studies were stable vital signs. In the studies of Sowter and Lecuru [18, 19] a HCG <5,000 IU/l was necessary. In the retrospective study of Hidlebaugh [22] and of Mol [20] no upper limit for the HCG was given. For Hidlebaugh et al., a sonographic diagnosis of EUG, or increasing HCG levels after dilatation and curettage fulfilled the inclusion criteria. For Mol et al. a diagnostic laparoscopy was essential for the diagnosis. Mol et al., managed EUGs with a HCG <1,500 IU/l expectative.

Alongside, the laboratory parameters, imaging with the transvaginal ultrasound (TVS) has tremendously improved the early diagnosis of the EUG. The sonographic detection of a fetal heartbeat was an exclusion criteria in four of the six studies [18–20, 22]. Also excluded were EUGs with a diameter >3.5 cm on ultrasound, and patients with sonographic free fluid (any or more than 300 ccm) as this was considered an indication for tubal rupture. Seror and Westaby did not publish their sonographic exclusion criteria [16, 21]. In all clinical studies, the excluded patients underwent surgical treatment.

The two medical protocols used were single-dose methotrexate (50 mg/m2) [18, 21, 22] and a multi-day alternating protocol (methotrexate im 1 mg/kg body weight on days 0, 2, 4, 6 and alternating folic acid rescue po 0.1 mg/kg) [19–21]. Westaby did adapt the protocol according to the HCG levels whilst Seror [16] did not provide details regarding the protocol.

As a possible bias must be considered the influence of the health systems on the study population, follow-up data, hospital and/or insurance costs. In articles reporting from nations with a national health system, for example UK or NZ [16, 18, 21] costs were calculated from an insurance or public sector point of view. In contrast, a study from a nation with private health care insurance (USA), costs were calculated for the in-/out-patient care and also for surgical/anaesthesiology care [22] from the hospital point of view. Unfortunately this study did not provide any details regarding the time or cost calculation basis. The other two studies [19, 20] provided specific details of the cost calculation basis (i.e. hourly wages, material costs).

A major problem of inter trial comparison is that results were reported in different currencies. In the earlier article [19] of Lecuru, the costs were transferred from French franc into US dollar (US$). But with the European currency union the calculation was done in Euros in the later publication [23]. Mol et al. [20] transferred the cost into US$ whilst other studies decided to keep their national currency [16, 18, 21]. As the economic value of each currency depends on various factors (i.e. inflation, stability of the national economy, tax, etc.) this makes it difficult to calculate a direct comparison. In this article, to eliminate these and other factors the difference in treatment costs is given in percentage of the total cost of the laparoscopic treatment.

Generally 40–60 % of the total cost relates to the inpatient stay making it the main expense [18, 21] In all studies, surgically treated patients are hospitalised longer (1.3–2.7 vs 0–1.1 days) than the medically treated patients. This explains the majority of differences in the treatment ‘arm expenses’. Further differences depend on the study protocol. For example as Hidlebough treated his medical group strictly as outpatients, inpatient time is given with 0 days. In studies with higher patient numbers, the average inpatient time for medically treated cases ranged from 0.68 to 1.05 days.

In medically managed treatment protocol, direct costs were mainly caused by the repeated laboratory controls and ultrasound examinations. In all studies patients were seen more often by the consultant if managed medically. The total cost for these appointments was also higher compared to the surgically managed cases (2.4–62/0.35–46). Mol et al. required a ‘diagnostic’ laparoscopy and Hidlebaugh et al. started the treatment with a dilatation and curettage. In both articles these costs are also considered [20, 22].

Discussion

The treatment change from laparotomy to laparoscopy and medical treatment is reflected in the publication of Hidlebaugh et al. [22]. From early on prospective studies evaluated the treatment costs. Depending on the health system the cost details vary. In the US health system, for example, a direct billing of the patient/insurance is common. Therefore more details regarding the direct costs (i.e. costs relating to inpatient stay, drugs, surgery or other material) were given by US authors. In nations with a national health service (i.e. UK), indirect costs (i.e. sick leave, outpatient appointments, travel costs) are more detailed as the health system might cover travel or hotel expenses. All reviewed articles except Hidlebaugh [22] do provide indirect cost details.

Cost-effectiveness

In all studies the medical treatment is at least equally if not more cost-effective. Mol et al. [20] reported higher costs for medical treatment. The authors critically note by replacing the diagnostic laparoscopy with non invasive tools (ultrasound/blood parameters); medical treatment would be equally cost-effective. The methotrexate protocol was used in an inpatient setting. With further experience in the drug, Westaby [21] investigated the medical protocol in an outpatient setting. Under these circumstances the costs of Mol et al. would be reduced by a further 1000 US$ (~18 %) and therefore clearly favour the medical treatment.

Depending on the article and definition of EUG this cost advantage was up to 90 % of the total cost [22]. Some authors expect a cost reduction by a more frequent use of ultrasound and therefore earlier diagnosis [16] or by guideline-adherent treatment [24].

Treatment complications

Another substantial cost contributor were treatment complications. The rate of complications ranged between 2.5 and 18 % [19, 21]. The most common complications were trophoblast persistence and the need for repeated methotrexate application, followed by vomiting, adverse drug reactions or severe bleeding. In case of a persisting pregnancy a repeat methotrexate injection can be necessary. The rate of re-injection ranged between 16 and 26 %. Of these cases 3.7–27 % still needed a surgical intervention after failed medical treatment. The cost of this surgery and resulting inpatient stay increased the cost for the medical treatment. The complication rate increased with the HCG levels. In cases with HCG >3,000 IU/l, the total medical treatment cost including the cost of complications was higher than the surgical cost [21]. Cost reductions were achieved in small EUGs (HCG <1,500/3,600 IU/l) in the prospective studies and the systematic reviews [16, 18–20, 22]. The ‘break even point’ was found to be with HCG levels of 1,500–3,000 IU/l. This corridor was also confirmed by the retrospective study of Westaby [21].

Ultrasound

No correlation was given for the sonographic findings. This might be due to the inter observer reliability of ultrasound examinations and/or the assumption that lab parameters are considered more reliable. Only Seror et al. [16] conclude that an increased use of diagnostic ultrasounds would help increase the rate of medically treated EUGs. But diagnostic criteria are not provided.

Follow-up

Further cost contributors were follow-up examinations. In five of the reviewed studies, the number of appointments was on average less than nine but the study of Seror [16] gives four times the number of outpatient contacts. This is due to the focus on the next pregnancy. Therefore this figure includes appointments beyond the treatment of the EUG. The end of the evaluation period was a pregnancy or three years post treatment. The cost for the follow-up is broken down to ultrasound appointments [19–21] or gynaecological assessments [18] showing the cost reduction in the surgical group compared to medical follow-up. Taking indirect costs into account (i.e. sick leave, travel costs,….) this sum increased the post-surgical total above the medical indirect cost [18, 19, 21].

The results of the retrospective studies [20–22] show that the rate of medically treated EUGs can be increased if treatment guidelines or recommendations are closely followed. The medical treatment of the EUG is more cost-efficient for small EUGs (HCG <1,500 IU/l) compared to laparoscopic intervention. In haemodynamically stable patients (HCG <1,500 IU/l), qualified TVS and repeated HCG observation can increase the rate of medical treatments.

With a HCG level between 1,500 and 3,000 IU/l the treatment costs depend on the calculation. In health systems with compensation for indirect costs the surgical treatment seems to be more cost-efficient. In a setting with a reduced/no compensation for follow-up, the medical treatment seems to be more cost-effective. With HCG levels >3,000 IU/l, the total cost of medical treatment is higher than the surgical treatment according to the reviewed literature. This is mainly due to the increase of cost intensive complications. Therefore medical treatment seems not to be cost-effective in EUGs with HCG levels >3,000 IU/l.

Depending on the health system and viewpoint cost savings between 50 and 90 % per case seem to be possible. All authors consent to the cost reduction due to medical treatment. These savings are reduced by an increase in follow-up appointments.

Conclusion

For small EUGs (defined as HCG <1,500 IU/l), the results favour the medical treatment option. With HCG <3,000 IU/l, the cost-effectiveness is unclear but seems to favour the medical treatment depending on the health system. Contributing to the follow-up cost are costs for medication, travel costs and sick leave. These are higher in the medical subgroup.