Granulomatous mastitis—a retrospective observational cohort analysis of treatment outcomes in over 600 patients

Background: Granulomatous mastitis (GM) is an uncommon inflammatory condition of the breast characterized by relapsing inflammation and sterile abscess formation. To date, there is little data on the natural history of the disease. With the lack of understanding of the etiology of this disease, there is also a lack of knowledge on how best to treat GM creating a difficult and frustrating problem for patients and the physicians who care for them. Several therapies are currently employed, including antibiotics, steroids, immune modulators such as methotrexate, antivirals, and surgery. There is a lack of studies comparing treatments and their effectiveness across a large number of patients. We aimed to use a large database to evaluate the effectiveness of treatment strategies across a large number of patients to see if we would determine a superior treatment approach. Methods: We utilized the anonymized data from the electronic health records of 105 million patients in 92 healthcare organizations (HCOs) collected through the TriNetX database to identify patients diagnosed with GM between 2015–2022. Outcomes measured included length of treatment time, number of treatments, and recurrence. Multivariate analysis was conducted and Python software was used to analyze differences among various treatments and their outcomes. Results: We identified 1,836 patients with a diagnosis of GM during that time. Within the treatment analysis logic, 669 patients were identified as having received treatment for GM during that time period. Patients ranged from age 9–89 years with an age mean value of 39±13.2 years and a median value of 37 [interquartile range (IQR), 30–44] years. It was most common in women, but 2.5% of cases were identified in men. The most common therapies were antibiotics (38.9%, 260/669), followed by steroids (22.3%, 149/669), followed by methotrexate (2.2%, 15/669). The most common combination of treatments was antibiotics + steroids (19.6%, 131/669). Kruskal-Wallis test did not show a difference between recurrence times for different treatment combinations (P=0.33) or between retreatment times for different combinations of treatments (P=0.30). This study showed no statistically significant difference in recurrence time or retreatment rate among different treatment types, regardless of the method of treatment employed. Conclusions: The recurrence and retreatment rates for GM were not significantly different among different treatment types or combinations of treatment types. This study suggests that among the patient population analyzed, there was no clearly superior treatment for GM and highlights the need for prospective studies to help identify the best course of treatment for this frustrating and complex condition.