Background: Management of pregnancy of unknown location (PUL) is still challenging. Early or inaccurate interventions may lead to termination of a normal intrauterine pregnancy, while an untreated ectopic pregnancy (EP) may lead to tubal rupture or even death. Objectives: To create a decision-making algorithm to determine probability of EP among PUL patients. Methods: In this cross-sectional study, we considered 522pregnant women with abdominal pain during their first trimester presenting to Imam Hossein Medical Center in Tehran between 2012 and 2018. The clinical signs, symptoms, medical histories, laboratory tests, and sonography of these patients were recorded in a questionnaire, and patients were divided in two groups of EP and non-EP. Finally, effective and significant factors were identified and entered into a decision-making tree for diagnosis of EP. Results: Patients divided into EP group including 188 women (36%) and Non-EP group of 334 women (64%). Of 92 variables measured for each patient, 41 variables had meaningful relationship with EP diagnosis (P-Value <0.05). Five Significant variables though six rules entered the decision-making tree including unilateral pain, tenderness, pelvic or abdominal free fluid, leukocytosis, and tachycardia. Conclusions: Clinicians can determine the probability of EP regarding to algorithm provided by the present study. In HCG positive and unilateral pain, EP probability is 95%, with logical consideration of laparoscopic intervention, and if positive HCG and bilateral pain is present along with abdominal tenderness, the probability of EP is 65%, and close observation of the patient might be considered. In bilateral pain, without abdominal tenderness, other characteristics including free fluid, leucocytosis, and tachycardia is used to determine probability of EP diagnosis.
