Can risk factors, clinical history and symptoms be used to predict the risk of ectopic pregnancy in women attending an early pregnancy assessment unit?

Abstract

OBJECTIVES: We aimed to examine whether risk factors and symptoms may be used to predict the likelihood of EP in women attending early pregnancy assessment units in the UK. METHODS: Prospective observational cohort study. Pregnant women under 12 weeks gestation were recruited in three London university hospitals between August 2012 and April 2013. One hospital continued the recruitment between January and June 2015. A standardized information sheet incorporating patient demographics, past medical history and symptoms was completed by patients and confirmed by examining clinicians. The outcome was final pregnancy location. RESULTS: There were 1320 eligible patients with a total of 72 EP (rate 5.5%). Pelvic pain and diarrhoea >3 times in the previous 24 hrs were independent symptoms that increased the risk of EP: relative risk 2.4 (95% CI, 1.4 to 4.0), and 2.2 (95% CI 1.08 to 4.5) respectively (p-values 0.002 and 0.03). The only other independent marker of risk was duration of bleeding; the risk of EP increased 20% (95% CI, 14% to 27%) for every one-day increment (p < 0.001). A logistic regression model incorporating these factors, demonstrated an area under the ROC curve of 0.73 (95% CI, 0.67 to 0.79). The prevalence of EP was low when there was no pelvic pain, no diarrhoea and the duration of bleeding ≤3 days: EP rate of 1.5% (6/391). When a single symptom was present, the EP rate increased to 5% (29/631), for two risk factors 8% (1/12) and three risk factors 9% (9/103), respectively. Women with pelvic pain and vaginal bleeding of any severity for >3 days had a high EP rate of 16% (23/146). In the nine women who also reported diarrhoea >3 times in the previous 24 hrs, two were EPs. CONCLUSIONS: Only the presence or absence of pain, presence or absence of diarrhoea >3 times in the previous 24 hrs, and duration of bleeding were symptoms that significantly increased the risk for EP in women attending early pregnancy units. Risk factors and symptoms alone could not be used to reliably predict women who have EP.