Search for a command to run...
Background Pyometra uteri (PU) is an uncommon gynecologic condition characterized by purulent accumulation within the uterine cavity. Although it is typically managed conservatively with drainage and antibiotics, severe cases - particularly those complicated by uterine rupture (UR) - are associated with substantial morbidity and high mortality. Evidence regarding factors associated with severe PU remains limited, especially in aging populations. This study, therefore, explored clinical and microbiological factors potentially associated with severe PU in elderly women. Methods This retrospective case-control study included 120 patients diagnosed with PU at a tertiary-care hospital in Japan between January 2014 and September 2024. Severe cases were defined as those requiring intensive care unit (ICU) or high care unit (HCU) admission. Clinical characteristics, microbiological findings, and outcomes were compared between severe (n=10) and non-severe (n=110) cases. Because of the limited number of severe cases, multivariable logistic regression was not feasible; therefore, univariate logistic regression analyses were performed, with additional age- and BMI-adjusted stratified analyses using the Mantel-Haenszel method. Receiver operating characteristic (ROC) curve analysis was used to explore the predictive performance of abscess size. Results The median age of the study population was 84 years. Ten patients (8.3%) were classified as severe; and all had UR. The in-hospital mortality rate among severe cases was 50%. Several clinical and microbiological variables were statistically associated with severe disease in univariate analyses, including fever, abdominal pain, presence of coexisting adenomyosis, detection of extended-spectrum β-lactamase (ESBL)-producing Escherichia coli in vaginal pus, positive urine culture, and increasing abscess size. ROC analysis suggested a possible abscess size threshold of 77 mm (area under the curve or AUC=0.77), with 60% sensitivity and 91% specificity for identifying severe cases. However, this cutoff was derived from the present dataset and requires external validation. Conclusions Systemic symptoms, larger abscess size, certain gynecologic comorbidities, and microbiological findings may be associated with severe PU. Given the small number of severe cases and the exploratory nature of the analyses, these findings should be interpreted cautiously and considered hypothesis-generating. Larger prospective studies are needed to validate these observations.