Does this woman have adnexal torsion?

Abstract

No questionnaire is currently available for the presurgical diagnosis of adnexal torsion (AT). Our objective was to develop a predictive model for AT, based on the Self Assessment Questionnaire for Gynecologic Emergencies (SAQ-GE) designed for triaging women with acute pelvic pain. We performed a multicenter prospective trial conducted in five hospitals in France. Four hundred and ninety-six (496) women with acute pelvic pain (Numeric Rating Scale>4), including 31 with AT, were recruited from September 2006 through April 2008. An AT score was built using the SAQ-GE. Five criteria were independently associated with AT confirmed by surgery: unilateral lumbar or abdominal pain [adjusted diagnostic odds ratio (aDOR), 23.3; 95% confidence interval (95% CI), 3.0-178]; absence of leucorrhea and metrorrhagia (aDOR, 7.0; 95% CI, 2.5-20), ovarian pain (aDOR, 5.5; 95% CI, 1.5-21), unbearable pain (aDOR, 5.0; 95% CI, 1.4-18) and vomiting (aDOR, 3.7; 95% CI, 1.6-9.0). The SAQ-GE torsion score was based on these five criteria and its values range from 0 to 10. The low-risk group (SAQ-GE torsion score<7), based on the score values, has a sensitivity (Se) of 96.7% (95% CI, 90.5-100), a negative predictive value of 99.7% (95% CI, 99.1-100) and a negative likelihood ratio (Lr-) of 0.05, ruling out AT with a probability of AT of 0.3% (95% CI, 0.0-0.9). Cross-validation of the model was performed using the jackknife resampling procedure, retrieving an unbiased Se of 87.1 (95% CI, 75.1-99.1) and a specificity of 74.2% (95% CI, 70.2-78.2). The SAQ-GE torsion score may prove useful for screening for AT in patients experiencing acute pelvic pain.