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

Introduction

Humans have been studying diseases of the breast since ancient times (1), with most of the focus on the diagnosis and treatment of breast cancer. Granulomatous mastitis (GM) is a condition that was only recently recognized as it was first described in 1972 by Kessler and Wolloch (2) as Idiopathic GM mimicking cancer, and it only became an International Classification of Diseases (ICD) code in 2015. It is a chronic inflammatory disease of the breast characterized by relapsing sterile draining breast abscesses and incidence has increased from 0% to 0.001% in non-Hispanics and 0.001% to 0.006% in the Hispanic population from 2016 to 2023 (3). Patients often present with palpable erythematous masses, spontaneously draining abscesses, and sinus tract formation (4). It can be painful and can affect any quadrant of the breast. The diagnosis can be difficult as it can mimic breast cancer on imaging (5).

Researchers are still trying to understand the underlying etiology of this condition, and many hypotheses have been proposed and are being studied. These include everything from autoimmune (4) to hormonal imbalances such as hyperprolactinemia (6), changes associated with birth and lactation (7,8), and more rare infectious causes such as Corynebacterium (9,10). Because we have yet to understand the etiology of this condition, the best course of treatment is still uncertain. There is an overall lack of understanding regarding how best to treat GM. Many treatment approaches have been suggested and adopted. The most common are observation (11,12), antibiotics (13), steroids (topical, intralesional, and systemic) (14,15), surgical interventions (including aspiration, incision and drainage, and excision) (14,16), and immune modulators (17). Regardless of the treatment, the risk of recurrence requiring retreatment is high.

At this time, there is no clear optimal treatment strategy, which can be frustrating for patients and the providers who care for patients with this uncommon condition. Most studies looking at GM treatment include fewer than 100 patients (18). The authors had previously used a large database to evaluate the incidence and prevalence of GM across various populations (3). We wanted to see if we could use this same data source to determine if any specific treatment or combination of treatments resulted in a lower rate of recurrence of symptoms requiring retreatment. We had hoped that looking at a larger population of patients with this condition could shed light on the optimal treatment strategies and guide further research into this condition. We present this article in accordance with the STROBE reporting checklist (available at https://abs.amegroups.com/article/view/10.21037/abs-25-26/rc).

Methods

This study adhered to the Declaration of Helsinki and its subsequent amendments. It utilized de-identified, aggregated data from TriNetX, and under U.S. Health Insurance Portability and Accountability Act (HIPAA) Privacy Rule §164.514(a), did not constitute human subjects’ research. Therefore, it was exempt from institutional review board oversight and the requirement for informed consent.

This is a retrospective observational review performed using a de-identified database. The data used in this study were collected on July 12, 2023, from the TriNetX Research Network (NLP enriched), which provided access to electronic medical records (diagnoses, procedures, medications, laboratory values, genomic information) from approximately 105 million patients in 82 healthcare organizations (HCOs) primarily in the USA.

All patients with a diagnosis of GM (ICD-10: N61.2) between January 1, 2015, and December 31, 2022, were queried. The timeline was chosen because GM was officially recognized with a separate ICD code in the 10th revision, implemented in 2015. The data was collected on July 12, 2023, thus we sought to evaluate data for the time period between when this became an official diagnosis code and the last calendar year before data collection. The data for the queried cohort was provided by 45 HCOs.

Out of this cohort we selected patients with the following events of interest: antimicrobials (VA AM000), steroids (RxNorm 8640, 6902, 1514, 3264, 8638, 10759, 5492, 2878), surgery procedures such as incision and drainage [current procedural terminology (CPT) 10060, 10061] and excision (CPT 19120, 19125, 19301), methotrexate (RxNorm 6851), and antivirals (RxNorm 73645, 281, 4678). We defined treatment time as 14 days for antibiotics, 30 days for steroids, 7 days for surgery, 182 days (6 months) for methotrexate, and 30 days for antiviral. If a treatment was performed, we included only those for whom a detracting diagnosis was not identified within seven days before the treatment. Detracting diagnoses for antibiotics included urinary tract infection (UTI) (N39.0, N30.0), pneumonia (J09–18), sexually transmitted disease (STD) (A51.0, A54–59, A63.8, A64, A70–N77), sinusitis (J01, J32), otitis (H60, H65, H66, H67), pharyngitis (J02.9), conjunctivitis (H10, H16.2), skin and soft tissue infection (L00–08) and hidradenitis (L73.2). Detracting diagnoses for steroids included Dermatitis (L20–30), bronchitis (J20, J40, J41), asthma (J45), psoriasis (L40), ulcerative colitis (K51), Crohn’s (K50), Lupus (M32, L93), arthritis (M05, M15–19) and ankylosing spondylitis (M45). Detracting diagnoses for methotrexate were ectopic pregnancy or molar pregnancy (O00–8), malignancy such as leukemia or lymphoma (C81–96), psoriasis (L40), and rheumatoid arthritis (M06.9). Detracting diagnosis for antiviral were herpes viral infections (B00.0–99), herpes zoster disease (B02.0–99), and varicella (B01.9, B01.89, B01.2). Thus, any patients who had a detracting diagnosis that corresponded to the treatment given within 7 days of receipt of the specific treatment for GM were excluded from the treatment analysis (Figure 1).

Figure 1 Analysis logic used for this study. Patients were queried through the TriNetX Research Network which provides access to electronic medical records from approximately 105 million patients from 82 health care organizations. Patients with a first diagnosis of GM between January 1, 2015 and December 31, 2022 were eligible. Patients with the events of interest within the following 60 days were selected including antibiotics, steroids, methotrexate, surgery, and antivirals and patients were excluded patients with a detracting diagnosis within 7 days before the treatment. Patients who received repeat treatment after the initial treatment period ended were identified as having re-treatment and patients who had another diagnosis of GM after the treatment period ended as having recurrence. GM, granulomatous mastitis; STD, sexually transmitted disease; UTI, urinary tract infection.

We conducted our analysis using a Python program (version 3.9), leveraging the pandas library (version 2.0.3) to evaluate treatment types, combinations, and durations. Treatments were classified as combined if one commenced before the previous one concluded. A treatment was considered finished if no prescription was issued within a predefined number of days specific to each type after the last recorded treatment. The length of treatment was calculated as the number of days between the first treatment record after the GM diagnosis and the treatment end. Additionally, we analyzed GM recurrence and re-treatment occurrence and time. We defined retreatment as any treatment records that occurred after the initial treatment was completed. However, the fact that retreatment occurred was unrelated to GM recurrence, and it could have happened regardless of whether GM recurrence occurred. We determined the end of treatment by looking for the absence of any treatment records within the specified time frame for each type of treatment. We identified recurrence as a GM diagnosis record that happened after the initial treatment ended. Our outcome search covered a 5-year period.

Statistical analysis

All statistical analyses were performed using Python software (version 3.9) with the pandas library (version 2.0.3). Continuous variables were presented as mean ± standard deviation and median with interquartile range (IQR). Categorical variables were reported as frequencies and percentages. For treatment duration, recurrence rates, and re-treatment rates, 95% confidence intervals were calculated using standard formulas for proportions and means. The Kruskal-Wallis test was used to compare time-to-event outcomes (recurrence time and re-treatment time) across different treatment groups, as the data did not meet assumptions for parametric testing. Statistical significance was set at P<0.05.

Results

Patient characteristics

We identified 1,684 patients with records of treatments after a diagnosis of GM and 669 patients were treated within the analysis logic without prior detracting diagnosis. The majority of patients were female, although 2.5% (17 out of 669) were men. GM was found in patients with an age mean value of 39±13.2 years and a median value of 37 (IQR, 30–44) years. Patients were primarily White (49%), with a significant portion of unknown race (29%) and a minor part of Black or African American patients (13%). We observed 44% of non-Hispanics or Latinos compared to 34% of Hispanics or Latinos and 22% of unknown ethnicity. GM was also six times more common in Hispanics or Latinos compared to non-Hispanic or Latinos (0.006% compared to 0.001%) (Table 1).

Treatment characteristics

More than half of patients (65%) were treated with only one type of treatment, 27% had a combination of two types of treatment, and 8% of patients underwent treatment with more than two types. The most common single-type treatment was antibiotics (38.9%), followed by steroids (22.3%), methotrexate (2.2%), surgery (1.3%), and antivirals (0.3%). The most common combination of treatments was antibiotics + steroids (19.6%) (Figure 2). The average length of treatment for the top three most frequent treatments, including combinations of treatments, was 17.6±8.3 days for antibiotics, 40.1±19.6 days for steroids, and 51.4±37 days for steroids + antibiotics combination (Figure 3).

Figure 2 Number of patients treated with individual and combinations of treatments. Patients were identified in the TriNetX database from 92 healthcare organizations from 2016–2022.

Figure 3 Treatment time by treatments and combinations. Patients were identified in the TriNetX database from 92 healthcare organizations from 2016–2022. This figure shows a box plot of treatment durations (in days, on a logarithmic scale) for granulomatous mastitis patients across different treatment options and combinations.

GM recurrence was observed in 44% of cases (310 patients). Recurrence time ranged from 1 to 1,396 days with a median of 21 (IQR, 7–86) days (Figure 4). Kruskal-Wallis test did not show a difference (P=0.33) between recurrence times for different treatment combinations. The highest recurrence rate for treatments with more than ten patients was 58% for steroids, followed by antibiotics and methotrexate with 47% each, and antibiotics + steroids with 44%. The lowest recurrence rate of 18% was observed for the combination of surgery + steroids (Figure 5).

Figure 4 Time to recurrence by treatment combinations. Patients were identified in the TriNetX database from 92 healthcare organizations from 2016–2022.

Figure 5 Recurrence rate by treatment combinations. Patients were identified in the TriNetX database from 92 healthcare organizations from 2016–2022.

Re-treatment was observed in 62% of cases (415 patients) with time ranging from 1 to 1,579 days [median: 41 (IQR, 14–137) days] (Figure 6). Kruskal-Wallis test did not show a difference (P=0.30) between re-treatment times for different treatment combinations. The highest re-treatment rate for treatments with more than ten patients was 68% for steroid treatment, followed by antibiotics + steroids with 66%. The lowest recurrence rate of 33% was observed for the combination of methotrexate + antibiotic + steroids (Figure 7 and Table 2).

Figure 6 Time to re-treatment by treatment combinations. Patients were identified in the TriNetX database from 92 healthcare organizations from 2016–2022.

Figure 7 Re-treatment rate by treatment combinations. Patients were identified in the TriNetX database from 92 healthcare organizations from 2016–2022.

Discussion

Our study analyzed the treatment outcomes for over 600 patients with GM, which is one of the largest collections of patients with a record of treatment for GM in the literature. The authors had previously used this database to evaluate the incidence and prevalence of GM over time and across different races and ethnicities and this data was published separately after performing a second data pull looking just at incidence and prevalence over a slightly longer time period (January 2015 to December 2023) (3) For this evaluation, we used the same source and the initial data gathered to compare outcomes from the most common treatment strategies individually and when combined. The treatment algorithms were much more complex and were not re-run with the additional 12 months of data. So the present study stopped in 2022. We found that there appears to be no apparent superior treatment. We could not identify a specific therapy or combination that resulted in a lower risk of recurrence of a diagnosis of GM or retreatment rates. This study included data from 45 Health Care Organizations mainly in the US but also internationally through the TriNetX database and was thus not limited to a single institution or regional pattern of pathology, diagnosis and treatment which aids in the generalizability toward a diverse population of patients with this condition.

Multiple small studies have looked at the comparative outcomes of treatment with one or two interventions over a small number of patients, some of which have shown promising results. Kafadir showed that 58% of patients in their study or 10 of the 17 patients included in their study who received a combination of methotrexate and oral steroids had a complete response in 2–3 months (19). In addition, a recent meta-analysis done by Sarmadian, which included this study, showed the lowest rate of recurrence with methotrexate and oral steroids at 4%. However, even in this meta-analysis of 65 studies looking at recurrence rates across treatment modalities only 7 studies included more than 100 patients and only one of those seven was randomized (19). The randomized study included in this meta-analysis, with over 100 patients, was performed by Cetin in 2019, with 124 patients showing no significant difference across treatments (20). Another meta-analysis published recently found different results with the lowest risk of recurrence with antibiotics and expectant treatment or observation when compared to surgery, steroids & immune modulating agents. However, this study acknowledged that the number of articles compared to other treatments in each category was small, with 5 studies with 53 observed patients and 12 studies with 178 patients treated with antibiotics (21). One of the larger randomized studies showing a trend toward a specific treatment was performed by Shabnam and was a randomized controlled trial of 82 patients comparing surgery alone or surgery followed by oral steroids for 6 months. This study found that the patients who had surgery followed by steroids trended toward a lower rate of recurrence (49.9% vs. 12.1%) but this was not statistically significant (22). Al-Jarrah showed that of their 20 patients treated with 6 weeks of antibiotics, all had complete remission with no recurrences over 15-month follow-up (23). However, the number of patients in most studies showing a benefit with a specific therapy is less than 100, and many of these studies showing benefits with a specific therapy have a study population of less than 10. Our study shows no specific superior treatment or combination of treatments over a much larger patient population, which is a key strength of this study.

This study is limited in that it was performed using a de-identified database and thus was only able to be performed in a retrospective observational fashion. This study design hinders the possibility of identifying causation, leaving room only for possible correlations and introducing the risks of confounding and biases. Additionally, this database relies on IDC-10 codes to query data and if data is missing or miscoded, this could affect analysis. The database also lacks information on pathology or imaging confirmation and there is no guarantee that the diagnosis of GM was clinically validated which may conflate the sample size. Another limitation is that the pertinent ICD-10 diagnosis codes were selected based on what our institution typically utilizes. Generally, the most popular steroids and antibiotics were selected, and a thorough list was created for each type of treatment, but these may differ by institution and provider. However, this list ought still to be typical of most institutions. We also did not include novel therapies not as widely described for treating GM in the literature. Also, given the way TriNetX shares data on patients, we lack information on the severity of the presentation of patients diagnosed with GM, which could impact how their physician decides to manage their treatment regimen.

Recently, an international multidisciplinary consensus statement evaluated multiple other studies looking at GM and recommended that patients with this condition be managed in a multidisciplinary setting to improve the overall outcome for these patients (24). As more groups move toward interdisciplinary care, this may allow the collaboration of multidisciplinary groups across institutions to study management strategies for this condition in a prospective manner across a larger patient population.

Our study highlights that we still have a long way to go in understanding the etiology and best way to manage GM and that even over a large patient population there was still no clear superior method of treatment. It also showed that even when comparing single treatments and combinations of treatments there was no regimen that had a lower recurrence or retreatment rate. Further prospective studies are needed to identify the best course of treatment for this frustrating and complex condition.

Conclusions

GM remains a very challenging condition to treat without a clear superior method of treatment to date. We were hopeful that use of a database to analyze a large number of patients might shed some light on how best to manage this condition. However, this study showed that among the patient data analyzed there does not appear to be a superior method of treatment or even combination of treatments that significantly reduced the recurrence or retreatment rate of GM when analyzed over a large population of patients with the condition. This highlights the need for further research on GM to determine the optimal treatment pathway and allow for improvement in outcomes for patients with this difficult diagnosis.

Acknowledgments

Our abstract has been presented as a poster presentation at the American Society of Breast Surgeons Annual Meeting in April 2024 in Orlando, FL, USA.

Footnote

Reporting Checklist: The authors have completed the STROBE reporting checklist. Available at https://abs.amegroups.com/article/view/10.21037/abs-25-26/rc

Data Sharing Statement: Available at https://abs.amegroups.com/article/view/10.21037/abs-25-26/dss

Peer Review File: Available at https://abs.amegroups.com/article/view/10.21037/abs-25-26/prf

Funding: None.

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://abs.amegroups.com/article/view/10.21037/abs-25-26/coif). V.S.K. serves as an unpaid editorial board member of Annals of Breast Surgery from March 2025 to December 2026. The other authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. This study adhered to the Declaration of Helsinki and its subsequent amendments. It utilized de-identified, aggregated data from TriNetX, and under U.S. HIPAA Privacy Rule §164.514(a), did not constitute human subject’s research. Therefore, it was exempt from institutional review board oversight and the requirement for informed consent.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Sudhakar A. History of Cancer, Ancient and Modern Treatment Methods. J Cancer Sci Ther 2009;1:1-4. [Crossref] [PubMed]
2. Kessler E, Wolloch Y. Granulomatous mastitis: a lesion clinically simulating carcinoma. Am J Clin Pathol 1972;58:642-6. [Crossref] [PubMed]
3. Zou J, Nelson N, Botnar K, et al. Clinical Trends in Granulomatous Mastitis Incidence, Prevalence, and Treatment: A Retrospective Study Highlighting Ethnic Differences in Care. J Surg Res 2024;302:732-8. [Crossref] [PubMed]
4. Yaghan R, Hamouri S, Ayoub NM, et al. A Proposal of a Clinically Based Classification for Idiopathic Granulomatous Mastitis. Asian Pac J Cancer Prev 2019;20:929-34. [Crossref] [PubMed]
5. Lee JH, Oh KK, Kim EK, et al. Radiologic and clinical features of idiopathic granulomatous lobular mastitis mimicking advanced breast cancer. Yonsei Med J 2006;47:78-84. [Crossref] [PubMed]
6. Yılmaz TU, Gürel B, Güler SA, et al. Scoring Idiopathic Granulomatous Mastitis: An Effective System for Predicting Recurrence? Eur J Breast Health 2018;14:112-6. [Crossref] [PubMed]
7. Brown KL, Tang PH. Postlactational tumoral granulomatous mastitis: a localized immune phenomenon. Am J Surg 1979;138:326-9. [Crossref] [PubMed]
8. Farouk O, Abdelkhalek M, Abdallah A, et al. Rifampicin for Idiopathic Granulomatous Lobular Mastitis: A Promising Alternative for Treatment. World J Surg 2017;41:1313-21. [Crossref] [PubMed]
9. Dobinson HC, Anderson TP, Chambers ST, et al. Antimicrobial Treatment Options for Granulomatous Mastitis Caused by Corynebacterium Species. J Clin Microbiol 2015;53:2895-9. [Crossref] [PubMed]
10. Johnstone KJ, Robson J, Cherian SG, et al. Cystic neutrophilic granulomatous mastitis associated with Corynebacterium including Corynebacterium kroppenstedtii. Pathology 2017;49:405-12. [Crossref] [PubMed]
11. Lai EC, Chan WC, Ma TK, et al. The role of conservative treatment in idiopathic granulomatous mastitis. Breast J 2005;11:454-6. [Crossref] [PubMed]
12. Yin Y, Liu X, Meng Q, et al. Idiopathic Granulomatous Mastitis: Etiology, Clinical Manifestation, Diagnosis and Treatment. J Invest Surg 2022;35:709-20. [Crossref] [PubMed]
13. Steuer AB, Stern MJ, Cobos G, et al. Clinical Characteristics and Medical Management of Idiopathic Granulomatous Mastitis. JAMA Dermatol 2020;156:460-4. [Crossref] [PubMed]
14. Lei X, Chen K, Zhu L, et al. Treatments for Idiopathic Granulomatous Mastitis: Systematic Review and Meta-Analysis. Breastfeed Med 2017;12:415-21. [Crossref] [PubMed]
15. Yildirim E, Kayadibi Y, Bektas S, et al. Comparison of the efficiency of systemic therapy and intralesional steroid administration in the treatment of idiopathic granulomatous Mastitis. The novel treatment for Granulomatous Mastitis. Ann Ital Chir 2021;92:234-41.
16. Yau FM, Macadam SA, Kuusk U, et al. The surgical management of granulomatous mastitis. Ann Plast Surg 2010;64:9-16. [Crossref] [PubMed]
17. Kehribar DY, Duran TI, Polat AK, et al. Effectiveness of Methotrexate in Idiopathic Granulomatous Mastitis Treatment. Am J Med Sci 2020;360:560-5. [Crossref] [PubMed]
18. Sarmadian R, Safi F, Sarmadian H, et al. Treatment modalities for granulomatous mastitis, seeking the most appropriate treatment with the least recurrence rate: a systematic review and meta-analysis. Eur J Med Res 2024;29:164. [Crossref] [PubMed]
19. Kafadar MT, Bahadır MV, Girgin S. Low-Dose Methotrexate Use in Idiopathic Granulomatous Mastitis: An Alternative Treatment Method. Breast Care (Basel) 2021;16:402-7. [Crossref] [PubMed]
20. Çetin K, Sıkar HE, Göret NE, et al. Comparison of Topical, Systemic, and Combined Therapy with Steroids on Idiopathic Granulomatous Mastitis: A Prospective Randomized Study. World J Surg 2019;43:2865-73. [Crossref] [PubMed]
21. Fattahi AS, Amini G, Sajedi F, et al. Factors Affecting Recurrence of Idiopathic Granulomatous Mastitis: A Systematic Review. Breast J 2023;2023:9947797. [Crossref] [PubMed]
22. Shabnam N, Choudhury MIM, Ahsan T. Treatment Modality of Idiopathic Granulomatous Mastitis with Surgery Alone or in Combination with Steroid. Bioresearch Communications 2021;7:924-6.
23. Al-Jarrah A, Taranikanti V, Lakhtakia R, et al. Idiopathic Granulomatous Mastitis: Diagnostic strategy and therapeutic implications in Omani patients. Sultan Qaboos Univ Med J 2013;13:241-7. [Crossref] [PubMed]
24. Yuan QQ, Xiao SY, Farouk O, et al. Management of granulomatous lobular mastitis: an international multidisciplinary consensus (2021 edition). Mil Med Res 2022;9:20. [Crossref] [PubMed]