The incidence of ectopic pregnancy (EP) has been reported to be increasing in the United States, but the representative rate is difficult to estimate because there have been decreased inpatient treatment and increased multiple healthcare surveillance visits. In Norway, the EP rate had a marked decline from 2000-2004, with rates similar to those observed in a study in 2005 that used computerized data systems from a large US managed care organization from 1997-2000. In general, medical management with methotrexate appears to have been increasing over time. Methotrexate therapy is a common alternative to surgery and has been proved to be effective, safe, and economical. There are various regimens of methotrexate administration that include the single-dose protocol, a fixed multiple-dose intramuscular regimen, and a variable multiple-dose regimen. The single-dose protocol is the most commonly used regimen, with reported success rates to be 65-96%. Successful treatment in this protocol is defined by a >15% decrease in beta–human chorionic gonadotropin (β-hCG) levels between days 4 and 7 after methotrexate administration. This 15% decrease from days 4-7 rule has been validated by previous investigators as an excellent indicator of treatment success, with reported positive predictive values up to 93.0%.

Potential disadvantages of the 15% decrease from days 4-7 rule include the need for patient compliance, the risk of treatment failure, and EP rupture while awaiting day 7 β-hCG levels. Patients are expected to return for repeat β-hCG titers at least 2 times, and many are often lost to follow up by day 7. Furthermore, patients with more “aggressive” EP because of persistent trophoblastic tissue may rupture by the time of the follow-up visit 1 week later. Therefore, an earlier predictor of response to methotrexate would be helpful in the clinical management of EP

The significance of early β-hCG changes after methotrexate therapy has been an area of interest among investigators as a predictor of treatment outcome after methotrexate. Rises in β-hCG titers have been observed in 50-70% of cases after methotrexate therapy and previously have not been correlated to the persistence of trophoblastic tissue. Early β-hCG values after single-dose methotrexate therapy has been combined with other modalities, such as transvaginal ultrasound (TVUS) findings in a logistic regression model, to predict the likelihood of tubal rupture. Declining β-hCG values of at least 20% between days 1 and 4 during methotrexate treatment have been associated with a positive predictive value of 97% for treatment success. In addition, absolute measurements of β-hCG values on day 4 after methotrexate injection have not been found to be predictive of success after single-dose therapy.

This study attempts to clarify the role of early β-hCG level changes as a predictor of persistent EP after methotrexate treatment for ectopic pregnancies and to determine the usefulness of β-hCG changes between days 0 and 4 as a predictor of methotrexate therapy success. We hypothesize that there exists a correlation between the rate of β-hCG increase between days 0 and 4 after a single-dose methotrexate injection and response to this therapy in EP; the higher rates of increase suggest the presence of more invasive trophoblastic tissue and therefore a higher chance of methotrexate failure.

Materials and Methods

Before this study began, institutional review board approval was obtained. Data were collected from the charts of women who had been diagnosed with an EP and treated with a single intramuscular dose of methotrexate (50 mg/m ² ) from January 2000 to January 2009 at Tampa General Hospital. The definitive diagnosis of an EP was made when patients had positive β-hCG titers that were associated with either (1) visualization of gestational sac outside of the uterus by TVUS or (2) the absence of trophoblastic tissue in the uterine cavity by anatomopathologic studies on curettage. A presumptive diagnosis of EP was made when patients had β-hCG levels above the discriminatory zone (1500-2000 IU/L) with the absence of an intrauterine pregnancy by TVUS or abnormally rising/plateauing β-hCG titers below the discriminatory zone that is associated with TVUS findings that are suggestive of EP (complex adnexal mass and/or free fluid in peritoneal cavity).

Recorded data included information from the history, physical, and laboratory values before and after treatment, specifically serum β-hCG levels on days 0, 4, and 7 after methotrexate therapy. Patient age, race, obstetric and gynecologic history, and TVUS findings were also recorded. Of note, the exclusion criteria for single-dose methotrexate therapy at our institution included hemodynamic instability, signs of an acute abdomen, abnormal baseline hematologic, renal, or hepatic laboratory values, adnexal mass >3.5 cm, and fetal cardiac activity.

Inclusion criteria for the study were complete medical records that fulfilled the aforementioned criteria for the diagnosis of EP and subjects with days 0, 4, and 7 β-hCG values appropriately recorded after methotrexate administration. Exclusion criteria included incomplete records, nonadherence to the protocol that was used at our institution ( Figure 1 ), extreme β-hCG values (>10,000 IU/mL), and patients who were lost to follow-up evaluation. Treatment success was defined as a >15% decrease in β-hCG levels between days 4 and 7, followed by weekly β-hCG level measurement until negative. Treatment failure was defined as the need for a second dose of methotrexate and/or surgery for EP removal.

Single-dose protocol at Tampa General Hospital

β-hCG , beta–human chorionic gonadotropin; CBC , complete blood count; CMP , complete metabolic panel; MTX , methotrexate; T&S , type and screen.

Nguyen. Early β-hCG changes after methotrexate therapy. Am J Obstet Gynecol 2010.

Statistical analysis was performed with the Fisher’s exact test for categoric variables and with t test or Mann-Whitney U test (where appropriate) for continuous variables. In addition, the difference variable (defined as the day-4 β-hCG value minus the day-0 β-hCG value) and the percentage change variable (defined as the day-4 β-hCG value minus the day-0 β-hCG value divided by the day-0 value multiplied by 100) were calculated for each case to investigate which data representation could best differentiate the groups that had treatment success and failure. A probability value of < .05 was considered statistically significant.