Aim: To investigate characteristics and risk factors for recurrent adnexal torsion (AT). Methods: Retrospective cohort study in a university-affiliated medical center included 320 Women with AT verified by laparoscopy, from January 2005 through January 2017. Demographic data, clinical symptoms, surgical findings and treatment were retrospectively reviewed from patient records. Data from primary and secondary AT of patients in the recurrent torsion group was compared to those with single torsion, to evaluate risk factors for recurrent AT. Results: Two hundred and sixty seven (83.4%) patients had a single event of AT and 53 (16.5%) had recurrent AT. Patients with recurrent torsion had significantly fewer previous non-gynecologic surgeries (4.3% and 9.8% of the study groups vs 32.2% of the control group, P = 0.001 for both). Ovarian size was significantly smaller in the recurrent torsion groups (47.5 mm and 48.3 mm vs 63.9 mm, P = 0.045 and P = 0.012, respectively). Polycystic ovary was significantly more common in the recurrent AT group (P = 0.028 and P = 0.005), with risk ratio 4.4 (95% confidence interval, 1.66 to 11.63). Ovaries without any specific findings were also more common among recurrent AT cases (P = 0.001 for both groups). Logistic regression analysis demonstrated that smaller ovarian size is an independent risk factor for recurrent AT. Conclusion: Recurrent torsion correlated with fewer previous surgeries, small ovarian mass, polycystic ovaries and ovaries without specific findings, which might indicate that additional pathophysiological factors contribute to the recurrent event. Ovarian fixation might be recommended in patients with primary torsion of normal or polycystic ovaries.