Surgical Intervention in Patients with Tubo-Ovarian Abscess: Clinical Predictors and a Simple Risk Score

Study Objective: To identify the clinical characteristics associated with surgical intervention in patients with tubo-ovarian abscess (TOA). Design: Retrospective cohort study (Canadian Task Force classification II-2). Setting: Tertiary university-affiliated hospital. Patients: Three hundred thirty-five patients were diagnosed with TOA based on sonographic and clinical criteria. Patients who underwent surgical intervention were compared with patients managed conservatively. Intervention: Electronic medical records were used to identify patients who were diagnosed with TOA between 2007 and 2015. All patients received the same antibiotic regimen upon admission. The data extracted included microbial and pathologic reports. Clinical characteristics such as disease severity and outcomes were compared. The clinical predictors available on patient admission for surgical intervention were identified retrospectively. A logistic regression was used to determine the independent predictors of treatment failure. A risk score was created by giving a nominal weight to each predictor. The score was validated by a random bootstrap analysis. An additional validation cohort that consisted of patients diagnosed with TOA during the 2 years after the original study period was applied to the final score. Measurements and Main Results: The following variables of patients who underwent surgical intervention in comparison with those successfully treated and were enrolled into the score analysis differed significantly: age at admission (odds ratio [OR], 2.1; 95% confidence interval [CI], 1.3–3.5), mean leukocytosis at admission (OR, 2.2; 95% CI, 1.3–3.6), ultrasonographic measurement of abscess diameter (OR, 3.6 95% CI, 2.0–6.3), and the presence of bilateral abscess (OR, 2.2; 95% CI, 1.3–3.9). Risk groups A, B, C, and D were positively correlated with the need for surgical intervention. Those in the highest risk group D had an antibiotic failure rate of 92%, as compared with those with the lowest risk group, in which there was a 20% risk of antibiotic failure. Conclusions: Antibiotic treatment failure for TOA can be predicted on patient admission using a novel risk assessment score.