Expectant management of tubal ectopic pregnancy: prediction of successful outcome using decision tree analysis

Abstract

Objective: To establish whether a decision tree based on a combination of clinical, morphological and biochemical parameters could be constructed to help in the selection of women with tubal ectopic pregnancies for expectant management.

Methods: This was a prospective observational study in a tertiary referral early pregnancy unit in an inner city teaching hospital. The study group consisted of 179 women with ultrasound diagnosis of ectopic pregnancy. Demographic, clinical and ultrasound data were recorded in each case at the initial visit. In addition all women had a blood sample taken for the measurements of serum beta-human chorionic gonadotropin (beta-hCG) and progesterone. Clinically stable women with non-viable pregnancies and no signs of hematoperitoneum were managed expectantly on an outpatient basis until their serum beta-hCG declined to <20 IU/L. Women who developed pelvic pain during follow-up and those with non-declining serum beta-hCG were offered surgery.

Results: A total of 107/179 (59.8%) tubal ectopics were considered suitable for expectant management. Ectopic pregnancy resolved spontaneously in 75/107 (70%) women, which was 41.9% of the total number of tubal ectopics. Maternal age, initial serum beta-hCG and progesterone were all significantly different in pregnancies that resolved spontaneously compared to those requiring surgery (P < 0.05). Initial serum beta-hCG level was the best predictor of the outcome of expectant management. These differences enabled a construction of a four-level decision tree to estimate the likelihood of successful expectant management.

Conclusions: There are significant differences in demographic, ultrasound and biochemical findings between spontaneously resolving ectopics and those requiring treatment. Decision tree analysis may be used as a guide to estimate the probability of successful expectant management in individual cases.