Since the early 1990s, more than half of the patients treated for unruptured tubal pregnancy in the authors' department have undergone unilateral linear salpingotomy. Although this procedure can preserve fertility, estimates of the frequency of persistent ectopic pregnancy (PEP) associated with salpingotomy range from 5% to 29%. This study was conducted to investigate various patient characteristics that might identify patients at greatest risk of developing PEP, including pre- and postprocedure serum levels of human chorionic gonadotropin (hCG). Data analyses were performed using the Student t test, Mann-Whitney U test, Fisher exact test, or Χ 2 test. From 1992 through 1998, 417 of 806 patients who were treated for ectopic pregnancy at Hvidovre Hospital underwent laparoscopic linear salpingotomy. In all, 48 women had PEP and required secondary treatment. Three women with persistent clinical symptoms were treated with salpingectomy on days 1, 7, and 46. Forty-five women were diagnosed with PEP because of rising or stagnant serum hCG levels. There were no clinical symptoms in 30 women. The median interval between salpingotomy and secondary treatment for these women was 15 days (range, 1-46 days). Eight patients were treated within 9 days of initial treatment, and 14 were treated within the first 12 days. In all, 23 of the 48 women with PEP chose to have salpingectomy, 23 chose systemic methotrexate, and 2 elected to undergo a second linear salpingotomy. Nineteen of the 23 patients treated with methotrexate received a multidose oral regimen. The regimen failed in four women, who then underwent salpingectomy. In 1998 the protocol for methotrexate was changed to a single-dose intramuscular injection. In this series the single dose was successful in all four women who received it. Nearly all of the women (94% of patients in whom salpingotomy was successful and 98% of those with PEP) had preoperative serum hCG values taken within 3 days of initial treatment. The preoperative values were significantly higher in patients who had PEP (median, 2100 IU/liter) than in successfully treated patients (median, 893 IU/liter; P = .003). A higher risk of PEP was associated with preoperative serum hCG levels of 2500 IU/liter and above; however, a receiver operating characteristic curve showed low prognostic accuracy for preoperative hCG levels at all cutoff points. Serum hCG levels from 2500 to 3000 IU/liter had a specificity for predicting PEP ranging from 0.74 to 0.77 and a sensitivity ranging from 0.38 to 0.45. Serum hCG levels were measured a median of three times for women who underwent successful linear salpingotomy and eight times for those with PEP. Of the 38 patients with PEP who had hCG levels measured on the ninth postoperative day, 25 had values 10% of preoperative levels, compared with 72 of 317 successfully treated women (P <.0001). Measurements taken on day 12 showed that 22 of 33 (66.7%) PEP patients had serum hCG values of 7% of initial values, compared with 80 of 324 successful cases (24.7%; P <.001). These two cutoff levels had a sensitivity and specificity of predicting PEP of 0.66 to 0.67 and 0.77 to 0.75, respectively. The prognostic value was poor at all cutoff levels. Multivariate analysis of potential predictive variables, including duration of surgery, surgical technique, history of previous ectopic pregnancy, history of previous lower abdominal surgery, ruptured ectopic pregnancy, pelvic adhesions, and absence of products of conception at microscopy and hemoperitoneum, found no predictive ability for any variable.
Read full abstract