OC068: Patient history and ultrasound findings: A logistic regression model for the prediction of pelvic pathology in women with chronic pelvic pain

To assess what factors in a patients history and which transvaginal ultrasound findings are most predictive of pathology in women with chronic pelvic pain (CPP). Prospective study. 120 consecutive women with CPP completed a questionnaire detailing their previous history and signs and symptoms. All patients had a transvaginal ultrasound scan (TVS) 1–2 weeks before laparoscopy. All patients had a negative pelvic infection screen. During the transvaginal scan (TVS), the vaginal probe was used to assess ovarian location and mobility, and site-specific tenderness. The average age of the patients was 31 years (range 18–50 years) and the average duration of pain was 46 months (range 6–180 months). At laparoscopy 70 (58%) patients had pelvic pathology and 50 (42%) patients had a normal pelvis. The historical factors—severity of symptoms (1–10), and previous gynaecological surgery were predictive of pelvic pathology (p 0.05). On TVS the presence of an endometrioma or hydrosalpinx were predictive of pelvic pathology (accuracy 100% and 86% respectively). Reduced ovarian mobility and site-specific tenderness were predictive of pelvic pathology (p < 0.01). Stepwise logistic regression revealed ultrasound based site-specific tenderness as the most predictive factor. Historical factors and TVS findings are considered important in the assessment of patients with CPP. In this study logistic regression analysis has shown site-specific tenderness on its own to be as accurate as all the other factors.