Adnexal torsion: a predictive score for pre-operative diagnosis.

BACKGROUND Adnexal torsion (AT) is difficult to diagnose and requires immediate surgery. The aim of this study was to develop a simple score for assisting in the pre-operative diagnosis of AT in women with acute pelvic pain. METHODS Using data from a retrospective cohort of 142 patients with acute pelvic pain, we developed a score based on multiple logistic regression after a jackknife procedure. We validated the score in a prospective cohort of 35 women with acute pelvic pain. RESULTS Five criteria were independently associated with AT confirmed by surgery: unilateral lumbar or abdominal pain [adjusted odds ratio (aOR), 4.1; 95% confidence interval (95% CI), 1.2-14.0]; pain duration <8 h at first presentation (aOR, 8.0; 95% CI, 1.7-37.5), vomiting (aOR, 7.9; 95% CI, 2.3-27.0), absence of leucorrhoea and metrorrhagia (aOR, 12.6; 95% CI, 2.3-67.6) and ovarian cyst larger than 5 cm by ultrasonography (aOR, 10.6; 95% CI, 2.9-38.8). The torsion score was based on these five criteria. Low-risk and high-risk groups were derived from values of the score [probability of AT, 3.7% (95% CI, 0-7.8) and 69% (95% CI, 53-84), respectively]. Application of these criteria to the prospective cohort confirmed the diagnostic accuracy of the score [probability of AT, 0% (95% CI, 0-16) and 75% (95% CI, 26-100) in the low-risk and high-risk groups, respectively]. CONCLUSIONS This easy-to-calculate score may prove useful for diagnosing AT in patients with acute pelvic pain seen at general or gynaecology emergency departments.