The paradoxical effect of oral prednisone in granulomatous lobular mastitis: An observational cohort study.

Skip to main content An official website of the United States government Here's how you know Here's how you know Official websites use .gov A .gov website belongs to an official government organization in the United States. Secure .gov websites use HTTPS A lock ( Lock Locked padlock icon ) or https:// means you've safely connected to the .gov website. Share sensitive information only on official, secure websites. Search Log in Dashboard Publications Account settings Log out Search… Search NCBI Primary site navigation Search Logged in as: Dashboard Publications Account settings Log in Search PMC Full-Text Archive Search in PMC Journal List User Guide PERMALINK Copy As a library, NLM provides access to scientific literature. Inclusion in an NLM database does not imply endorsement of, or agreement with, the contents by NLM or the National Institutes of Health. Learn more: PMC Disclaimer | PMC Copyright Notice PLoS One . 2026 Apr 8;21(4):e0341901. doi: 10.1371/journal.pone.0341901 Search in PMC Search in PubMed View in NLM Catalog Add to search The paradoxical effect of oral prednisone in granulomatous lobular mastitis: An observational cohort study Haiyan Zhang Haiyan Zhang 1 Breast Centre, West China Second University Hospital, Chengdu, China Conceptualization, Formal analysis, Supervision, Validation, Writing – original draft, Writing – review & editing Find articles by Haiyan Zhang 1 , Ruiyang Wu Ruiyang Wu 2 Department of Breast and Thyroid, Sichuan Provincial Hospital for Women and Children (Affiliated Women and Children’s Hospital of Chengdu Medical College), Chengdu, China Data curation, Formal analysis, Methodology, Validation, Writing – original draft Find articles by Ruiyang Wu 2 , Jin Chen Jin Chen 2 Department of Breast and Thyroid, Sichuan Provincial Hospital for Women and Children (Affiliated Women and Children’s Hospital of Chengdu Medical College), Chengdu, China Data curation, Formal analysis, Methodology, Writing – original draft Find articles by Jin Chen 2 , Jinyan Feng Jinyan Feng 1 Breast Centre, West China Second University Hospital, Chengdu, China Investigation, Methodology Find articles by Jinyan Feng 1 , Jing Luo Jing Luo 1 Breast Centre, West China Second University Hospital, Chengdu, China Conceptualization, Investigation, Supervision, Writing – review & editing Find articles by Jing Luo 1, * Editor: Siddharth Gosavi 3 Author information Article notes Copyright and License information 1 Breast Centre, West China Second University Hospital, Chengdu, China 2 Department of Breast and Thyroid, Sichuan Provincial Hospital for Women and Children (Affiliated Women and Children’s Hospital of Chengdu Medical College), Chengdu, China 3 Sai Gosavi Specialty Clinic / Nano Hospitals Bangalore / Saraswati Specialty Clinic, INDIA Competing Interests: The authors have declared that no competing interests exist. ✉ * E-mail: [email protected] Roles Haiyan Zhang : Conceptualization, Formal analysis, Supervision, Validation, Writing – original draft, Writing – review & editing Ruiyang Wu : Data curation, Formal analysis, Methodology, Validation, Writing – original draft Jin Chen : Data curation, Formal analysis, Methodology, Writing – original draft Jinyan Feng : Investigation, Methodology Jing Luo : Conceptualization, Investigation, Supervision, Writing – review & editing Siddharth Gosavi : Editor Received 2025 Oct 28; Accepted 2026 Jan 14; Collection date 2026. © 2026 Zhang et al This is an open access article distributed under the terms of the Creative Commons Attribution License , which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. PMC Copyright notice PMCID: PMC13061257  PMID: 41950240 Abstract Background Despite being a cornerstone therapy for granulomatous lobular mastitis (GLM), the use of oral prednisone lacks a standardized protocol due to undefined optimal parameters and insufficient large-scale evidence regarding its impact on disease recurrence. Objectives The retrospective cohort study comprehensively assessed the efficacy of oral prednisone for GLM by utilizing a large-scale cohort. It aimed to identify the optimal therapeutic protocol, with the ultimate aim of establishing a robust evidence base to inform standardized clinical guidelines. Methods A cohort of 614 patients diagnosed with GLM was included in this study from January 2017 to December 2024. A three-dimensional evaluation of treatment timing (symptom-to-treatment interval), duration, and initial dosage (both absolute and weight-adjusted) was performed to identify prognostic factors and optimal cutoffs. Results Oral prednisone was identified as an independent risk factor for recurrence (HR = 2.534, 95% CI: 1.615–3.975, P < 0.001). Survival curve analysis revealed that the 1-year recurrence rate was significantly higher in the oral prednisone group than in the non-prednisone group (29.4% vs. 14.6%, P = 0.001). Subgroup analysis showed that the long-interval, low-dose group (symptom-to-treatment interval > 6 weeks, dosage ≤ 0.4 mg/kg/day) yielded the most favorable results. The one-year recurrence rate of this group (6.2%) was significantly lower than other groups (all P < 0.05) and comparable to the non-prednisone group (14.6%, P = 0.233). Conversely, the short-interval group (symptom-to-treatment interval: 0–6 weeks) demonstrated the highest one-year relapse rate (33.5%), significantly exceeding that of the non-prednisone group (P < 0.001). Conclusions While oral prednisone therapy may be associated with an increased relapse risk in GLM, an optimized protocol—initiating treatment after a 6-week symptom interval at a dose of ≤ 0.4 mg/kg/day—was identified to mitigate this risk. These findings highlight the paramount importance of treatment timing and dosage, offering a data-driven strategy to refine current therapeutic practices. Prospective studies are warranted to validate this proposed regimen. Introduction Granulomatous lobular mastitis (GLM) is a rare, benign, and chronic inflammatory disease of the breast with an unknown etiology [ 1 , 2 ]. Nonetheless, emerging evidence from large-scale, population-based studies indicates a rising diagnostic frequency of GLM in recent years, suggesting it is transitioning from a rare entity to a more commonly encountered clinical condition [ 3 ]. GLM typically presents with rapid-onset inflammation that progresses to abscesses and chronic, disfiguring sinus tracts, which often results in a severe and protracted disease course persisting for one to two years [ 4 , 5 ]. There is no universally agreed-upon therapeutic strategy for GLM, and its management is actively debated. Nonetheless, corticosteroid-based regimens have established themselves as a cornerstone of treatment due to their profound anti-inflammatory and immunomodulatory effects [ 6 – 9 ]. Corticosteroid therapy for GLM encompasses both systemic and topical approaches [ 10 ]. Although local steroid therapy (e.g., intralesional injection) has been reported as a useful treatment modality [ 6 , 11 ] and offers the potential advantages of direct drug delivery and reduced systemic exposure compared to oral therapy, its widespread clinical adoption remains limited by the need for professional administration, injection-related discomfort, and a lack of large-scale comparative evidence. Systemic steroid therapy, most commonly employing prednisone, is more frequently utilized [ 7 ]. However, this approach faces several major challenges. First, recurrence rates following systemic corticosteroid treatment are notably high, with reported figures ranging widely from approximately 25% to over 45% across different studies [ 4 , 7 , 12 , 13 ]. Second, a proportion of patients experience adverse effects—such as weight gain, hirsutism, hyperglycemia, and iatrogenic cushing—which frequently impair treatment compliance and limited therapy duration [ 14 ]. Finally, owing to the limited evidence from small case series [ 1 , 4 , 7 , 12 – 17 ], key therapeutic parameters—including the optimal starting dose, treatment timing, and total treatment duration—have not been established, leaving clinicians without a standardized, evidence-based regimen. Given the lack of consensus and the highly variable practices in systemic corticosteroid therapy, this study comprehensively assessed the efficacy of oral prednisone for GLM by utilizing a large-scale cohort. The aim was to identify the optimal therapeutic protocol, with the ultimate goal of establishing a robust evidence base to inform standardized clinical guidelines. Materials and methods Data and samples We conducted a large-scale cohort study using data from a proprietary non-puerperal mastitis database established in March 2022 at a major tertiary maternity hospital. The database enrolled patients who were diagnosed with non-puerperal mastitis (NPM) according to the Chinese expert consensus [ 18 ] who have been treated at the institution since January 2017. Prospective data collection commenced in March 2022, while data prior to this date were collected retrospectively. Data collection followed a standardized protocol throughout both phases. Prospective data were recorded in real time using predefined forms, while retrospective data were systematically re-extracted from original records by two independent investigators applying the same variable definitions. This process was guided by an internal database manual, with institutional details anonymized for review. The diagnosis of GLM for all included cases was established through a comprehensive assessment, integrating clinical presentation and imaging findings, and was definitively confirmed by histopathological examination. Specifically, GLM cases were identified and confirmed based on pathological reports, after the exclusion of infectious etiologies (e.g., tuberculosis) and other specific granulomatous diseases, consistent with expert diagnostic pathways [ 19 ]. Subsequently, to focus the analysis on the efficacy of oral prednisone, the following three study-specific exclusion criteria were applied to the initial GLM cohort: (1) patients diagnosed with plasma cell mastitis (a distinct NPM subtype); (2) patients with missing body weight records (which precluded calculation of weight‑adjusted doses); and (3) patients who received corticosteroid therapy other than oral prednisone. From an extraction on September 10, 2025, we identified 710 patients diagnosed between January 2017 and December 2024. After applying exclusion criteria, we sequentially excluded 44 patients with plasma cell mastitis, 8 with unknown body weight, and 44 who received non-prednisone corticosteroid therapy, resulting in a final analytical cohort of 614 patients with GLM. The protocol for this cohort study, which utilized a private database, was approved by the Institutional Ethics Committee in March 2022 (Approval No. 20220330−024) prior to its initiation. The study was conducted in compliance with the ethical principles of the Declaration of Helsinki (2013 revision). As the research involved only the analysis of pre-existing, anonymized data without any patient intervention, the ethics committee granted a waiver of informed consent. This study defined recurrence as the appearance of NPM symptoms (such as redness, swelling, pain, or a palpable mass) in the ipsilateral or contralateral breast following documented initial clinical improvement or cure. Cure was defined as complete resolution of symptoms, disappearance of palpable masses, and negative findings on both physical examination and imaging studies [ 18 ]. All potential recurrence events during follow-up were identified and adjudicated by the study's database manager, a senior breast disease specialist, applying the predefined criteria above. The time-to-event analysis measured the duration from diagnosis to recurrence or the final follow-up, with all durations rounded to the nearest month. The final follow-up date was September 1, 2025. The laboratory parameters analyzed in this study (white blood cell count, C-reactive protein, prolactin) were obtained from multiple tests conducted throughout the patients’ clinical course. Given that not all patients underwent testing at initial presentation, and to establish a consistent measure of disease activity, the following approach was adopted: the highest recorded value from all available test results for each patient, spanning from their first visit for the index episode to the last follow-up, was extracted and analyzed. This method is based on the clinical rationale that the peak level of inflammation or hormonal activity during the disease course may better reflect its maximum biological intensity and was used for subsequent prognostic analysis. Hyperlipidemia was defined as meeting at least one of the following: total cholesterol ≥ 220 mg/dL, low-density lipoprotein cholesterol ≥ 140 mg/dL, high-density lipoprotein cholesterol < 40 mg/dL, or triglycerides ≥ 150 mg/dL [ 20 ]. This study employed two main treatment modalities: pharmacological therapy and surgical intervention. The pharmacological strategies encompassed a range of antimicrobials (fluoroquinolones, penicillins, cephalosporins, macrolides, nitroimidazoles), anti-tuberculosis agents (isoniazid, rifampin, or ethambutol), oral prednisone (the regimen, including the starting dose and tapering schedule, was determined by the attending physician based on clinical judgment, commonly involving an initial dose followed by a taper of 5 mg every 3–7 days until discontinuation), and bromocriptine. Surgical interventions involved procedures for abscess drainage (e.g., aspiration, vacuum-assisted drainage, open debridement) and surgical excision (via vacuum-assisted biopsy or open surgery). Statistical analysis Continuous variables were expressed as mean and standard deviation (SD), while categorical variables were presented as frequencies and percentages. To investigate the impact of oral prednisone treatment on relapse in patients with GLM, two analytical methods were employed. First, a Cox proportional hazards regression model was used to identify potential predictors of disease relapse, with a focus on oral prednisone therapy. Second, patients were divided into two groups: the oral prednisone group (patients who received oral prednisone at the entire course of the disease) and non-prednisone group (patients who never received any corticosteroid therapy at the entire course of the disease, including oral prednisone or local treatment). Propensity score matching (PSM) was used to balance covariates that differed between the two groups. The survival curves of the two groups before and after matching were compared using the Breslow test to evaluate the treatment effect. We further analyzed the prognostic role of oral prednisone through three dimensions: (1) treatment timing, by stratifying patients into short-interval and long-interval groups based on the time from symptom onset to treatment initiation; (2) treatment duration, by categorizing patients into short-duration and long-duration groups according to the total course of therapy (excluding maintenance doses); and (3) treatment dosage (excluding maintenance doses). To comprehensively evaluate the dose-effect relationship, the dosage was analyzed using two complementary metrics: the absolute initial dose (mg/day) and the weight-adjusted initial dose (mg/kg/day). The absolute dose reflects the overall intensity of drug exposure, whereas the weight-adjusted dose accounts for inter-patient physiological differences, providing a more personalized and clinically relevant pharmacologic measure. PSM was performed using the MatchIt package in R [ 21 ], implementing a 1:1 nearest-neighbor algorithm with a caliper width of 0.1. The optimal cutoff values for the above groups were identified using X-tile software (version 3.6.1, Yale University School of Medicine, New Haven, CT, USA) [ 22 ], which applies the minimum P-value method from survival analysis. This X-tile analysis was considered hypothesis-generating; all suggested thresholds were independently validated via time-trend and multi-dimensional subgroup analyses. Based on this risk stratification, subgroup analyses were conducted to identify the optimal oral prednisone strategy (detailed in Fig 1 ). All hypothesis tests were two-sided, with a P-value of less than 0.05 considered statistically significant. The statistical analyses were performed using R software (version 4.2.2; R Foundation for Statistical Computing, Vienna, Austria), and the survival curves were plotted using Stata (version 19; StataCorp LLC, College Station, TX, USA). Fig 1. Research workflow of this study. Open in a new tab Results Descriptive characteristics and prognostic factor analysis This study enrolled a total of 614 GLM patients with a mean age of 31.4 ± 4.8 years, and all participants were female. The mean follow-up duration was 13.9 ± 14.9 months. The one-year recurrence rate was 23.9%. Univariable and multivariable COX regression analyses ( Table 1 ) identified the following factors as independent predictors of recurrence: affected side (P < 0.001), white blood cell count (P < 0.001), antitubercular therapy (HR = 0.394, 95% CI: 0.264–0.588, P < 0.001), oral prednisone (HR = 2.534, 95% CI: 1.615–3.975, P < 0.001), and abscess drainage (HR = 0.506, 95% CI: 0.300–0.854, P = 0.011). Table 1. Univariable and multivariate analysis of prognostic factors for 614 patients with granulomatous lobular mastitis. Factors N(%)/x ± s Univariate Multivariate P value HR(95%CI) P value Age at diagnosis (year) 31.4 ± 4.8 0.126 0.940(0.905-0.977) 0.002 Weight (kg) 60.5 ± 10.3 0.292 1.012(0.995-1.029) 0.171 Affected side <0.001 <0.001 left 319(52.0) Ref right 284(46.3) 1.987(1.402-2.816) <0.001 bilateral 11(1.8) 11.275(4.479-28.380) <0.001 Days to first visit (days) 19.1 ± 28.5 0.093 0.985(0.977-0.993) <0.001 Maximum lesion diameter on ultrasound 0.059 0.021 < 4 cm 239(38.9) Ref ≥ 4 cm 337(54.9) 1.611(1.081-2.400) 0.019 unknown 38(6.2) 2.407(1.029-5.629) 0.043 Ultrasound-detected lesion count 0.806 0.190 solitary 97(15.8) Ref multiple 478(77.9) 0.928(0.558-1.542) 0.772 unknown 39(6.4) 0.415(0.159-1.078) 0.071 Mammary abscess 0.271 0.710 no 101(16.4) Ref yes 513(83.6) 1.120(0.618-2.030) Microabscess 0.087 0.094 no 234(38.1) Ref yes 243(39.6) 0.653(0.436-0.978) 0.039 unknown 137(22.3) 0.978(0.583-1.641) 0.933 White blood cell <0.001 <0.001 < 10*10^9/L 336(54.7) Ref ≥ 10*10^9/L 259(42.2) 1.784(1.205-2.641) 0.004 unknown 19(3.1) 5.516(2.508-12.134) <0.001 C-reactive protein 0.096 0.021 < 10 mg/L 221(34.4) Ref ≥ 10 mg/L 155(25.2) 1.023(0.623-1.681) 0.928 unknown 248(40.4) 1.755(1.121-2.748) 0.014 Prolactin <0.001 0.070 ≤ 650 uIU/ml 391(63.7) Ref > 650 uIU/ml 156(25.4) 0.922(0.575-1.478) 0.735 unknown 67(10.9) 1.909(1.072-3.398) 0.028 Hyperlipidemia 0.590 0.905 no 194(31.6) Ref yes 194(31.6) 0.961(0.619-1.493) 0.861 unknown 226(36.8) 1.068(0.681-1.675) 0.776 Quinolone therapy 0.596 0.248 no 145(23.6) Ref yes 469(76.4) 1.313(0.827-2.087) Penicillin therapy 0.847 0.674 no 592(96.4) Ref yes 22(3.6) 0.798(0.280-2.279) Cephalosporin therapy 0.893 0.124 no 456(74.3) Ref yes 158(25.7) 1.368(0.918-2.039) Macrolide therapy 0.874 0.449 no 586(95.4) Ref yes 28(4.6) 1.341(0.628-2.865) Nitroimidazole therapy 0.985 0.221 no 589(95.9) Ref yes 25(4.1) 1.722(0.721-4.111) Antitubercular therapy <0.001 <0.001 no 377(61.4) Ref yes 237(38.6) 0.394(0.264-0.588) Oral prednisone <0.001 <0.001 no 245(39.9) Ref yes 369(60.1) 2.534(1.615-3.975) Bromocriptine therapy 0.371 0.553 no 344(56.0) Ref yes 270(44.0) 1.137(0.745-1.735) Abscess drainage 0.005 0.011 no 186(30.3) Ref yes 428(69.7) 0.506(0.300-0.854) Surgical excision 0.074 0.022 no 128(20.8) Ref yes 486(79.2) 0.585(0.369-0.927) Open in a new tab Abbreviations: HR, hazard ratio; CI, confidence interval; Ref, reference. The normal reference range for serum prolactin is 80.56–650.84 uIU/ml. To evaluate the robustness of our findings with respect to the data collection mode, we conducted a sensitivity analysis using only the prospectively collected data from March 2022 onward (n = 374). This approach aimed to exclude potential influences arising from variations in completeness or consistency inherent in the earlier retrospective records. Univariable and multivariable Cox regression analyses performed on this prospective cohort (see S1 Table ) yielded results highly consistent with the primary analysis. Affected side (P < 0.001), white blood cell count (P = 0.001), antitubercular therapy (HR = 0.433, 95% CI: 0.225–0.833, P = 0.012), oral prednisone (HR = 2.645, 95% CI: 1.332–5.253, P = 0.005), and abscess drainage (HR = 0.439, 95% CI: 0.229–0.839, P = 0.013) remained independent predictors of recurrence. The direction of effect (protective or risk) and the magnitude of the hazard ratios for these predictors were closely aligned with those obtained from the primary analysis of the entire cohort (n = 614). Oral prednisone After stratifying the 614 patients into an oral prednisone group (n = 369) and a non-prednisone group (n = 245) based on treatment exposure, survival analysis before PSM revealed a significantly higher 1-year recurrence rate in the oral prednisone group (29.4% vs. 14.6%, P = 0.001, Fig 2A ). To address potential confounding, 1:1 PSM was performed, balancing all baseline characteristics (e.g., lesion count, abscess presence, white blood cell count, and prior treatments) between 171 matched pairs (all P > 0.05 after matching, S2 Table ). In this matched cohort, the oral prednisone group continued to show a significantly elevated 1-year recurrence rate compared to the non-prednisone group (21.8% vs. 13.6%, P = 0.042, Fig 2B ). These results indicate that oral prednisone exposure is independently associated with an increased risk of recurrence, even after accounting for baseline differences through rigorous matching. Fig 2. A: Comparison of 1-year recurrence rates: oral prednisone versus non-prednisone groups before propensity score matching; B: Comparison of 1-year recurrence rates: oral prednisone versus non-prednisone groups after propensity score matching. Open in a new tab Treatment timing To further investigate the impact of the timing of oral prednisone therapy on disease recurrence, we assessed the association between the time interval from symptom onset to treatment initiation and clinical outcomes. Of the 369 eligible patients, 335 with complete data on the symptom-to-treatment interval (34 missing, 9.2%; mean: 5.2 weeks; SD: 7.1 weeks) were analyzed. Using X-tile software, the optimal cutoff for this interval was identified as 6 weeks. Stratifying the cohort accordingly, we found that patients in the short-interval group (≤ 6 weeks, n = 253) had a 1-year recurrence rate of 33.5%, which was significantly higher than the 15.7% observed in the long-interval group (> 6 weeks, n = 82) (P = 0.016, Fig 3A ). The time-varying recurrence analysis further confirmed the 6-week cutoff, showing that the 1-year recurrence rate peaked (range: 29.1%−38.4%) when treatment started within 6 weeks but plummeted to 15.1% and 14.4% for initiations at 7–8 and ≥ 9 weeks, respectively ( Fig 3B ). Fig 3. A: Kaplan-meier survival curves for the short-interval and long-interval groups; B: Sharp decline in 1-year recurrence rate with treatment delay beyond 6 weeks. Open in a new tab Treatment duration We assessed the effect of oral prednisone treatment duration (excluding maintenance doses) on outcomes in a cohort of 341 patients with available data (from an original cohort of 369; 28 excluded due to missing exposure data, 7.6%). The mean treatment duration was 3.5 weeks (SD: 2.7 weeks). X-tile software determined 4 weeks as the optimal cutoff for recurrence risk stratification. Stratification of the cohort based on this cutoff revealed no significant difference in the 1-year recurrence rate between the short-duration (< 4 weeks, n = 215) and long-duration (≥ 4 weeks, n = 126) groups (30.0% vs. 30.5%, P = 0.381, Fig 4A ). Time-dependent analysis ( Fig 4B ) further validated that the 1-year recurrence rate was similar across various treatment durations, ranging from 23.7% (0–1 week) to 33.0% (≥ 6 weeks), which suggests that the length of oral prednisone course (excluding maintenance) was not a major determinant of recurrence. Fig 4. A: Kaplan-meier survival curves for the short-duration and long-duration groups; B: The 1-year recurrence rate by oral prednisone duration. Open in a new tab Treatment dosage The investigation of the impact of the initial dose of oral prednisone on prognosis comprised two parts: the absolute initial dose (mg/day) and the weight-adjusted dose (mg/kg/day). First, we analyzed the absolute initial dose. From the original cohort, 337 patients with complete records were included (32 excluded due to missing data, 8.7%). The mean absolute dosage of oral prednisone was 24.7 ± 7.7 mg. Using X-tile software, the optimal cutoff was determined to be 20 mg/day. Patients were stratified into low-absolute-dose group (≤ 20 mg/day, n = 163) and high-absolute-dose group (> 20 mg/day, n = 174). The analysis revealed a significantly lower 1-year recurrence rate in the low-absolute-dose group compared to the high-absolute-dose group (24.4% vs. 35.6%, P = 0.031, Fig 5A ). Time-dependent analysis ( Fig 5B ) revealed a critical threshold: recurrence rates showed a notable jump starting at the 25–30 mg/day dosage (33.1%), with the sharpest increase observed in the ≥ 35 mg/day group (43.2%). This is in contrast to the lower rates seen at ≤ 10 mg/day (26.8%) and 15–20 mg/day (24.1%), highlighting a significantly elevated risk beyond 25 mg/day. Fig 5. A: Kaplan-meier survival curves for the low-absolute-dose and high-absolute-dose groups; B: Time-dependent analysis of 1-year recurrence rate by absolute initial oral prednisone dosage. Open in a new tab Second, we assessed the impact of the weight-adjusted dose (mg/kg/day) of oral prednisone on prognosis. From the original cohort, 337 patients with complete records were included in this analysis (32 were excluded due to missing data, 8.7%). The mean weight-adjusted dose was 0.41 ± 0.12 mg/kg/day. Using X-tile software, the optimal cutoff value was determined to be 0.43 mg/kg/day. Patients were stratified into low weight-adjusted dose group (≤ 0.43 mg/kg/day, n = 202) and high weight-adjusted dose group (> 0.43 mg/kg/day, n = 135). The analysis showed that the low weight-adjusted dose group had a significantly lower 1-year recurrence rate compared to the high weight-adjusted dose group (25.4% vs. 37.3%, P = 0.038, Fig 6A ). The 337 patients were categorized into five groups according to the weight-adjusted prednisone dose (mg/kg/day), as follows ( Fig 6B ): 0.2 mg/kg/day (≤ 0.24), 0.3 mg/kg/day (0.25–0.34), 0.4 mg/kg/day (0.35–0.44), 0.5 mg/kg/day (0.45–0.54), and 0.6 mg/kg/day (≥ 0.55). Analysis of the dose-response relationship revealed a non-linear trend in the one-year relapse rate. The rate was lowest (22.5%) at 0.3 mg/kg/day, increased modestly at 0.4 mg/kg/day (27.1%), and then rose sharply to 38.9% at 0.5 mg/kg/day, remaining elevated at 0.6 mg/kg/day (34.3%). Since the risk of relapse increased significantly beyond the 0.4 mg/kg/day dose, this level was proposed as the optimal cutoff value for maximizing efficacy and minimizing risk. Fig 6. A: Kaplan-meier survival curves for the low weight-adjusted dose and high weight-adjusted dose groups; B: Identification of a critical risk threshold at 0.4 mg/kg/day for weight-adjusted oral prednisone dose using time-dependent analysis. Open in a new tab Subgroup analysis Based on these findings, we analyzed 306 patients who received oral prednisone treatment. The patients were initially stratified into Groups 1–8 ( Table 2 ) according to the symptom-to-treatment interval, weight-adjusted dose (mg/kg/day), and treatment duration (63 patients were excluded due to missing essential grouping data, from an original cohort of 369 patients). The weight-adjusted dose (mg/kg/day) was selected over the absolute initial dose (mg/day) because corticosteroid dosing is typically adjusted according to body weight in clinical practice. Survival analysis showed that among patients with a symptom-to-treatment interval of 0–6 weeks (Groups 1–4), the one-year relapse rates ranged from 30.9% to 37.6%, with no statistically significant differences between these groups (all P > 0.05). However, Groups 1–4 had significantly higher relapse rates than Group 5 (3.8%, all P < 0.05, Table 2 ). Therefore, patients from Groups 1–4 were merged with 22 additional patients who had a symptom-to-treatment interval of 0–6 weeks but lacked complete data on dose or treatment duration, forming the short-interval group. Among patients with a symptom-to-treatment interval of 0–6 weeks and a dosage > 0.4 mg/kg/day (Groups 7 and 8), the one-year relapse rates were comparable (26.9% vs. 28.6%, P = 0.289, Table 2 ). Thus, Groups 7 and 8 were combined with 1 other patient meeting the same clinical criteria (symptom-to-treatment interval of 0–6 weeks and dosage > 0.4 mg/kg/day) but with missing treatment duration data, to constitute short interval-high dose group. For patients with a symptom-to-treatment interval of 0–6 weeks and a dosage ≤ 0.4 mg/kg/day (Groups 5 and 6), the one-year relapse rates were also not significantly different (3.5% vs. 11.1%, P = 1.000, Table 2 ). Accordingly, short interval-low dose group was formed by merging Group 5, Group 6, and an additional 2 patients who had a symptom-to-treatment interval of 0–6 weeks and a dosage ≤ 0.4 mg/kg/day, but whose treatment duration data were missing. Table 2. The difference in 1-year recurrence rates among groups 1-8. Group Symptom-to-treatment interval (weeks) Dosage (mg/kg/day) Duration (weeks) N 1-year recurrence rate (%) P value P value P value P value P value P value P value 1 0-6 ≤0.4 0-3 60 34.5 Ref 2 0-6 ≤0.4 ≥4 53 34.8 0.705 Ref 3 0-6 >0.4 0-3 84 37.6 0.521 0.328 Ref 4 0-6 >0.4 ≥4 34 30.9 0.756 0.928 0.408 Ref 5 ≥7 ≤0.4 0-3 27 3.8 0.009 0.030 0.004 0.026 Ref 6 ≥7 ≤0.4 ≥4 12 11.1 0.075 0.145 0.049 0.114 1.000 Ref 7 ≥7 >0.4 0-3 23 26.9 0.348 0.669 0.176 0.556 0.033 0.125 Ref 8 ≥7 >0.4 ≥4 13 28.6 0.125 0.217 0.079 0.149 0.694 0.609 0.289 Open in a new tab Abbreviations: Ref, reference. To further investigate the impact of oral prednisone administration on prognosis, a subgroup analysis was performed. The analysis revealed a clear hierarchy in relapse risk: the short-interval group had the highest 1-year relapse rate (33.5%), which was significantly greater than that of the non-prednisone group (14.6%, P < 0.001, Table 3 , Fig 7 ). In contrast, the relapse rate in the long interval-high dose group (25.9%) was comparable to that of the short interval group (P = 0.103, Table 3 , Fig 7 ). Notably, the long interval-low dose group exhibited the most favorable outcome, with a relapse rate (6.2%) that was not only significantly lower than other dosing groups (all P < 0.05, Table 3 , Fig 7 ) but also comparable to the non-prednisone group (P = 0.233, Table 3 , Fig 7 ). These results underscore that a short dosing interval is associated with the highest risk of relapse. Table 3. Comparison of 1-year recurrence rates among subgroups. Group Symptom-to-treatment interval (weeks) Dosage (mg/kg/day) Duration (weeks) N 1-year recurrence rate (%) P value P value P value non-prednisone group — 245 14.6 Ref short-interval group 0-6 no restriction no restriction 253 33.5 <0.001 Ref long interval-high dose group ≥7 >0.4 no restriction 37 25.9 0.467 0.103 Ref long interval-low dose group ≥7 ≤0.4 no restriction 41 6.2 0.233 0.002 0.038 Open in a new tab Abbreviations: Ref, reference. Fig 7. Kaplan-meier survival curves for subgroups. Open in a new tab Discussion Since its initial description in 1972 [ 23 ], no universally accepted treatment consensus has been established for GLM. The most widely accepted hypothesis posits GLM as an autoimmune process, which forms the pathophysiological rationale for employing corticosteroid therapy [ 24 ]. Studies have indicated that an initial course of oral steroids can significantly reduce breast and skin lesions, thereby potentially avoiding repeated and deforming operations [ 8 ]. To our knowledge, no study had previously provided a detailed, comparative analysis of oral prednisone treatment protocols for GLM, a critical knowledge gap this research sought to address. Based on a relatively large cohort, this study observed that treatment with oral prednisone was associated with a significantly increased risk of recurrence in patients with GLM (HR: 2.534), which corresponded to an approximately 1.5-fold increase in recurrence risk. Although initial use of oral steroids could lead to noticeable regression of breast and skin lesions and facilitate more limited surgery [ 8 , 25 ], our findings suggested that this did not necessarily lead to a reduction in long-term recurrence rates. Several lines of evidence supported this conclusion. First, the 1-year recurrence rate with oral prednisone therapy in our cohort was 29.4%, a result that aligned with the wide range of rates (24.7% to 46.9%) reported in earlier studies [ 4 , 7 , 12 , 13 , 16 , 26 , 1 ]. In addition, while corticosteroid therapy was found to suppress inflammation and reduce lesion size by inhibiting pro-inflammatory cytokines, it seldom led to complete disease eradication [ 12 ]. The high recurrence rate following oral steroid therapy was thus thought to be due to the persistence of subclinical residual foci. Furthermore, several studies reported a higher recurrence rate with oral steroids compared with observation alone [ 16 , 26 , 27 ], as illustrated in a 2024 meta-analysis of 65 studies where recurrence rates were 24% and 11%, respectively [ 26 ]. These consistent findings across studies, including ours, strengthen the observed association, although its interpretation must consider the potential for confounding inherent in observational data. This association remained robust even after rigorously controlling for baseline confounders via PSM, which yielded a significantly higher recurrence rate in the matched prednisone group (21.8% vs. 13.6%). A notable aspect of this study was its sample size. To our knowledge, the cohort included more than 600 patients, which exceeded the size of most previous related studies [ 1 , 4 , 7 , 12 – 17 , 25 , 28 – 32 ] (typically involving ≤ 200 patients), thereby offering a more robust foundation for the evidence in this area, though causal inference is constrained by the non-randomized study design. Although oral prednisone was widely used for GLM, the optimal timing for its initiation remained unclear. In this context, our findings highlighted the potential importance of treatment timing: a noticeable association with recurrence risk was observed, as patients treated within 6 weeks of onset experienced a recurrence rate exceeding 30%, compared to approximately 15% when treatment was initiated later ( Fig 3A ). In interpreting the observed association between early treatment initiation and higher recurrence, it is crucial to first consider a key alternative explanation: confounding by disease severity. Patients who presented and required treatment within the first 6 weeks likely had more active or severe disease at baseline than those who presented later. Thus, the higher recurrence in the early-treatment group may substantially reflect this greater intrinsic disease burden rather than a direct, negative effect of early treatment per se. Residual confounding by unmeasured severity factors cannot be entirely excluded and represents a primary limitation in causal inference. Even considering this predominant confounding effect, the interplay between the natural history of GLM and the immunomodulatory action of prednisone may offer complementary insights. First, if early immunosuppression during the acute phase achieves only temporary control without resolving underlying immune dysregulation, recurrence upon dose reduction (“steroid tapering rebound”) may be more likely. Second, the 6-week interval may act as a natural filter, allowing mild, self-limiting cases to resolve, thereby selecting a subgroup more amenable to systemic therapy. Third, delayed administration during a more chronic phase might modulate tissue repair more effectively, reducing pro-inflammatory scar formation that could serve as a nidus for recurrence. Collectively, these findings suggested that for GLM patients, there may be a preferable therapeutic window for oral prednisone—initiating treatment beyond 6 weeks, rather than very early, might be associated with improved long-term outcomes. This association requires further validation in studies designed to better control for baseline disease severity. To our knowledge, this study was among the first to systematically explore the timing of corticosteroid therapy in GLM, offering insights that could inform clinical practice. No unified standard currently exists for the course of oral prednisone, and clinical practice often relies on empirical medication. This study found that the length of the oral prednisone course (excluding the maintenance phase) was not a major determinant of recurrence, a finding contrasting with reports from other study favoring longer fixed courses [ 25 ]. Several factors may explain these discrepancies. Firstly, the differential requirements for prednisone course length based on disease severity and treatment response—with longer courses needed for severe or refractory cases and shorter ones sufficient for mild or rapidly responsive cases—may have led to the inconsistent study results. Thus, treatment duration may act more as a marker of disease behavior than as an independent modifiable factor. Moreover, the specific treatment regimens likely differed in key aspects beyond just course length. Notably, the initial dose in our cohort (0.41 mg/kg/d) was substantially lower than the 0.75 mg/kg/d used in the study advocating longer courses [ 25 ], suggesting dose intensity may be more critical than duration alone. Finally, the treatment context is paramount. The preferred strategy combines preoperative steroids with surgery [ 8 , 33 ]. In contrast, prednisone monotherapy, while a necessary alternative for non-surgical candidates, is associated with higher recurrence risk and often requires prolonged therapeutic courses [ 16 , 26 , 27 ]. Our study included a subset of patients who did not undergo surgery. These patients likely needed extended steroid courses yet still experienced high recurrence rates. This observation—that prolonged monotherapy signifies a refractory, high-risk subgroup—explains why the highest recurrence rate (33.0%) in our cohort was found in patients with treatment courses ≥ 6 weeks. Consequently, an isolated comparison of “treatment duration” lacks informativeness without critical context, particularly regarding concomitant surgical resection. In this context, a key contribution of our study is its provision of quantitative evidence from a large-scale cohort supporting the potential efficacy of a lower-dose regimen. Future work is necessary to establish standardized protocols that define the role and timing of surgery in conjunction with steroid therapy for more definitive clinical guidance. The dosage of oral prednisone was considered theoretically important for determining prognosis in GLM. However, the management of GLM with corticosteroids continued to be guided largely by clinical experience. Although current guidelines recommended a prednisone dose of 0.75 mg/kg/d [ 18 ], this full dosage was seldom achieved in clinical practice or research due to its magnitude. Importantly, few studies evaluated the impact of different dosages on clinical outcomes in GLM patients, leaving a notable evidence gap worldwide. In this study, which sought to address this question, a clear dose-outcome relationship was identified. The relapse risk increased notably at doses exceeding 0.4 mg/kg/d, suggesting that ≤ 0.4 mg/kg/day may serve as a reasonable therapeutic cutoff for balancing efficacy and safety. Crucially, when interpreting this relationship, an important consideration is that the prescribed dose may be correlated with disease severity at baseline (i.e., confounding by indication). It is plausible that patients with milder disease were prescribed lower doses, which could partly explain their better outcomes. As this study did not formally compare clinical severity between dose groups, this possibility cannot be ruled out and represents a key consideration for the results, also applying to the weight-adjusted dose analysis. Several previous studies supported the use of lower-dose prednisolone, reporting favorable outcomes and reduced recurrence rates [ 29 , 31 ]. One study demonstrated that a regimen of low-dose prednisolone (30 mg/day) combined with drainage was more effective and resulted in a lower recurrence rate than either surgical resection alone or high-dose prednisolone (50 mg) [ 29 ]. Similarly, Shin et al. [ 31 ] reported that initiating treatment with a steroid dose of 20 mg/day, in combination with surgery, was associated with a recurrence rate of only 7.1%. Together, these findings suggested that lower-dose corticosteroid regimens, particularly when integrated with surgical intervention, may help reduce recurrence. However, an alternative perspective was provided by Montazer et al. [ 32 ], who reported that a high-dose prednisone regimen (50 mg/day) was associated with better success rates and lower recurrence compared to a low-dose regimen (5 mg/day), potentially reducing the need for surgery. Despite these contrasting findings, the results of the present study were supported by several considerations. First, the analysis was based on a relatively large cohort [ 1 , 7 , 12 – 17 , 25 , 28 – 32 ], which helped improve the reliability of the conclusions. Nevertheless, the potential confounding effect of disease severity on dose assignment persists. Second, the rationale for a lower-dose strategy aligned with the principle of using the minimal effective dose to avoid excessive immunosuppression—a consideration especially relevant given the self-limiting nature of some GLM cases. This approach was also consistent with the Chinese expert consensus, which recommended an initial methylprednisolone dose of 20 mg [ 18 , 19 ]. Third, higher-dose regimens were often associated with more pronounced side effects, which could affect treatment adherence and thus influence outcomes [ 14 ], further supporting the exploration of an optimal, lower-dose strategy. To our knowledge, this was among the first studies to systematically examine the relationship between prednisone dosage and clinical outcomes in GLM. Addressing this evidence gap represented a useful step toward more informed treatment guidance. It is important to contextualize these findings within the observational nature of our study. The specific protocols for initiating and tapering prednisone were not uniform but reflected prevailing clinical practice. The absence of standardized definitions for an “initial dose” and a consensus tapering schedule highlights a significant gap in current GLM management. Future research must move toward prospective designs that define and test specific treatment protocols to establish a robust evidence base for optimal steroid use in GLM. Based on subgroup analyses, a clear hierarchy in relapse risk emerged. The long-interval, low-dose group had the most favorable outcome (1-year relapse rate: 6.2%), comparable to the non-prednisone group. In contrast, groups initiating treatment within 6 weeks had consistently high relapse rates (30.9%−37.6%). Integrating these findings, we propose the following clinical guidance to optimize the use of oral prednisone in GLM: (1) Timing and dose are paramount. If prednisone is used, initiating therapy beyond 6 weeks after symptom onset at a lower dose (≤ 0.4 mg/kg/day) is associated with the lowest relapse risk. (2) Avoid unnecessary exposure. For patients with small lesions and no marked systemic inflammatory response, the use of prednisone should generally be avoided. (3) Prioritize surgery in complex cases. For patients presenting with complex abscesses or fistulas, the cornerstone of management should be timely surgical intervention. Prednisone, if used, should play only a limited adjunctive role. It should be noted that these subgroup analyses regarding treatment timing, dosage, and duration are exploratory in nature. The compared subgroups (e.g., short- vs. long-interval, low- vs. high-dose) were not randomized and may have inherent baseline differences. Therefore, these findings, while informative and hypothesis-generating, require further validation in future prospective, ideally randomized, studies to confirm the proposed optimal treatment window and dosing strategy. This study has several limitations that should be acknowledged. First, its single-center, observational design may limit generalizability and introduces the potential for unmeasured confounding (e.g., confounding by indication for prednisone use), despite adjustments for multiple severity indicators. Specifically, while propensity score matching was employed to balance observed confounders including concomitant therapies, it cannot account for unmeasured or residual confounding factors. Furthermore, the reduction in sample size after matching may affect statistical power and the generalizability of the matched cohort. Second, although we applied a standardized protocol, the hybrid retrospective-prospective design over a long timeframe (2017−2024) carries an inherent risk of subtle inconsistencies in data or practice. Third, the analyses of specific prednisone parameters (timing, dose, duration) were exploratory and based on complete-case data; the missingness, though low (7.6%−9.2%), was not imputed. Fourth, and most pertinent to the primary findings regarding treatment parameters, the optimal thresholds for timing and dose were identified using the data-driven X-tile software. Although these thresholds were subsequently validated via time-trend analysis, multi-dimensional subgroup analysis, and multivariable modeling, the inherent risk of overfitting from such methods cannot be entirely excluded. These findings provide important hypotheses for optimizing prednisone strategies, and the precise cutoff values warrant external validation in prospective studies. Finally, this study focused on isolated prednisone use; defining its optimal role in combination with surgery remains a critical question for future research. Conclusions The findings suggested that oral prednisone might be associated with increased relapse rates in GLM patients. The analysis indicated that the symptom-to-treatment interval appeared to be the most influential factor in treatment outcomes, followed by weight-adjusted dosage, while treatment duration did not show a significant effect. This supports a cautious approach to prednisone use in GLM management. Based on these observations, a treatment approach involving prednisone initiation after a 6-week symptom interval, at a dose of ≤ 0.4 mg/kg/day, may be considered. Further studies will be needed to validate these preliminary findings and refine the proposed strategy. Supporting information S1 Table. Univariable and multivariable Cox regression analysis restricted to the prospective cohort (n = 374). (DOCX) pone.0341901.s001.docx (27.2KB, docx) S2 Table. Baseline characteristics of patients before and after propensity score matching. (DOCX) pone.0341901.s002.docx (28KB, docx) S1 Checklist. PLOSOne clinical studies checklist. (DOCX) pone.0341901.s003.docx (34.1KB, docx) Data Availability The datasets generated and analyzed in this study are not publicly available as they belong to a proprietary clinical database. Access can be granted for research purposes upon formal request. Interested researchers should contact the institutional data steward, Director Hu (Department of Breast and Thyroid Surgery), at [email protected] . Applicants were asked to specify the intended use of the data and to commit to compliance with the terms of use. Any data sharing was contingent upon final approval from the relevant institutions and the signing of a legally binding data use agreement. Funding Statement The author(s) received no specific funding for this work. References 1. Wang J, Zhang Y, Lu X, Xi C, Yu K, Gao R, et al. Idiopathic Granulomatous Mastitis with Skin Rupture: A Retrospective Cohort Study of 200 Patients Who Underwent Surgical and Nonsurgical Treatment. J Invest Surg. 2021;34(7):810–5. doi: 10.1080/08941939.2019.1696905 [ DOI ] [ PubMed ] [ Google Scholar ] 2. Cui L, Sun C, Guo J, Zhang X, Liu S. Pathological manifestations of granulomatous lobular mastitis. Front Med (Lausanne). 2024;11:1326587. doi: 10.3389/fmed.2024.1326587 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 3. Zou J, Nelson N, Botnar K, Khanipov K, Klimberg VS. Clinical Trends in Granulomatous Mastitis Incidence, Prevalence, and Treatment: A Retrospective Study Highlighting Ethnic Differences in Care. J Surg Res. 2024;302:732–8. doi: 10.1016/j.jss.2024.08.001 [ DOI ] [ PubMed ] [ Google Scholar ] 4. Tan Q-W, Zhang Y-N, Jia Y-P, Gou J, Lv Q, Yang X-Q. Methylprednisolone for idiopathic granulomatous mastitis: a prospective observational cohort study. Gland Surg. 2022;11(9):1538–45. doi: 10.21037/gs-22-484 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 5. Ocal K, Dag A, Turkmenoglu O, Kara T, Seyit H, Konca K. Granulomatous mastitis: clinical, pathological features, and management. Breast J. 2010;16(2):176–82. doi: 10.1111/j.1524-4741.2009.00879.x [ DOI ] [ PubMed ] [ Google Scholar ] 6. Zhou Y, Xu L. Clinical efficacy of different methods for treatment of granulomatous lobular mastitis: A systematic review and network meta-analysis. PLoS One. 2025;20(2):e0318236. doi: 10.1371/journal.pone.0318236 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 7. Toktas O, Konca C, Trabulus DC, Soyder A, Koksal H, Karanlik H, et al. A Novel First-Line Treatment Alternative for Noncomplicated Idiopathic Granulomatous Mastitis: Combined İntralesional Steroid İnjection with Topical Steroid Administration. Breast Care (Basel). 2021;16(2):181–7. doi: 10.1159/000507951 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 8. Gurleyik G, Aktekin A, Aker F, Karagulle H, Saglamc A. Medical and surgical treatment of idiopathic granulomatous lobular mastitis: a benign inflammatory disease mimicking invasive carcinoma. J Breast Cancer. 2012;15(1):119–23. doi: 10.4048/jbc.2012.15.1.119 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 9. Akahane K, Tsunoda N, Kato M, Noda S, Shimoyama Y, Ishigakis S, et al. Therapeutic strategy for granulomatous lobular mastitis: a clinicopathological study of 12 patients. Nagoya J Med Sci. 2013;75(3–4):193–200. [ PMC free article ] [ PubMed ] [ Google Scholar ] 10. Ma X, Min X, Yao C. Different Treatments for Granulomatous Lobular Mastitis: A Systematic Review and Meta-Analysis. Breast Care (Basel). 2020;15(1):60–6. doi: 10.1159/000501498 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 11. Gunduz Y, Altintoprak F, Tatli Ayhan L, Kivilcim T, Celebi F. Effect of topical steroid treatment on idiopathic granulomatous mastitis: clinical and radiologic evaluation. Breast J. 2014;20(6):586–91. doi: 10.1111/tbj.12335 [ DOI ] [ PubMed ] [ Google Scholar ] 12. Çetin K, Sıkar HE, Feratoğlu F, Taşdoğan B, Güllüoğlu BM. Treatment of Granulomatous Mastitis With Steroids: Should the Decision to End the Treatment be Made Radiologically?. Eur J Breast Health. 2023;20(1):25–30. doi: 10.4274/ejbh.galenos.2023.2023-9-2 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 13. Sarkar DK, Banerjee R, Gupta S, Singhal AK, Halder A. Management of idiopathic granulomatous mastitis: a prospective study. Ann R Coll Surg Engl. 2023;105(3):218–24. doi: 10.1308/rcsann.2022.0017 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 14. Çetin K, Sıkar HE, Göret NE, Rona G, Barışık NÖ, Küçük HF, et al. Comparison of Topical, Systemic, and Combined Therapy with Steroids on Idiopathic Granulomatous Mastitis: A Prospective Randomized Study. World J Surg. 2019;43(11):2865–73. doi: 10.1007/s00268-019-05084-x [ DOI ] [ PubMed ] [ Google Scholar ] 15. Tian C, Han X, Liu Z, Lv X, Ning P. Methotrexate and low-dose corticosteroid: An effective alternate against corticosteroid-resistant granulomatous lobular mastitis. J Obstet Gynaecol Res. 2022;48(11):2956–63. doi: 10.1111/jog.15396 [ DOI ] [ PubMed ] [ Google Scholar ] 16. Hur SM, Cho DH, Lee SK, Choi M-Y, Bae SY, Koo MY, et al. Experience of treatment of patients with granulomatous lobular mastitis. J Korean Surg Soc. 2013;85(1):1–6. doi: 10.4174/jkss.2013.85.1.1 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 17. Alper F, Karadeniz E, Güven F, Çankaya BY, Yalcin A, Özden K, et al. Comparison of the Efficacy of Systemic Versus Local Steroid Treatment in Idiopathic Granulomatous Mastitis: A Cohort Study. J Surg Res. 2022;278:86–92. doi: 10.1016/j.jss.2022.04.015 [ DOI ] [ PubMed ] [ Google Scholar ] 18. Zhou F, Liu L, Yu ZG. Expert consensus on the diagnosis and treatment of non-puerperal mastitis. Chinese Journal of Practical Surgery. 2016;36(7):755–8. [ Google Scholar ] 19. Yuan Q-Q, Xiao S-Y, Farouk O, Du Y-T, Sheybani F, Tan QT, et al. Management of granulomatous lobular mastitis: an international multidisciplinary consensus (2021 edition). Mil Med Res. 2022;9(1):20. doi: 10.1186/s40779-022-00380-5 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 20. Seishima M. Guideline for hyperlipidemia. Rinsho Byori. 2003;51(6):576–80. [ PubMed ] [ Google Scholar ] 21. Ho D, Imai K, King G, Stuart EA. MatchIt: Nonparametric Preprocessing for Parametric Causal Inference. J Stat Softw. 2011;42:1–28. [ Google Scholar ] 22. Camp RL, Dolled-Filhart M, Rimm DL. X-tile: a new bio-informatics tool for biomarker assessment and outcome-based cut-point optimization. Clin Cancer Res. 2004;10(21):7252–9. doi: 10.1158/1078-0432.CCR-04-0713 [ DOI ] [ PubMed ] [ Google Scholar ] 23. Kessler E, Wolloch Y. Granulomatous mastitis: a lesion clinically simulating carcinoma. Am J Clin Pathol. 1972;58(6):642–6. doi: 10.1093/ajcp/58.6.642 [ DOI ] [ PubMed ] [ Google Scholar ] 24. Ertürk TF, Çakır Ö, Yaprak Bayrak B, Güneş A, Aydemir S, Utkan NZ. Local Steroid Treatment: An Effective Procedure for Idiopathic Granulomatous Mastitis, Including Complicated Cases. J Invest Surg. 2022;35(4):745–51. doi: 10.1080/08941939.2021.1933272 [ DOI ] [ PubMed ] [ Google Scholar ] 25. Wang P, Sun J-Z, Fang H-Y, Yang D-J, Ren G-S. Optimal Timing for Corticosteroid Therapy in Idiopathic Granulomatous Mastitis: A Retrospective Analysis Highlighting Early Intervention Efficacy. J Inflamm Res. 2024;17:9617–24. doi: 10.2147/JIR.S498018 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 26. Sarmadian R, Safi F, Sarmadian H, Shokrpour M, Almasi-Hashiani A. 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(1):164. doi: 10.1186/s40001-024-01761-3 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 27. Lei X, Chen K, Zhu L, Song E, Su F, Li S. Treatments for Idiopathic Granulomatous Mastitis: Systematic Review and Meta-Analysis. Breastfeed Med. 2017;12(7):415–21. doi: 10.1089/bfm.2017.0030 [ DOI ] [ PubMed ] [ Google Scholar ] 28. Yabanoğlu H, Çolakoğlu T, Belli S, Aytac HO, Bolat FA, Pourbagher A, et al. A Comparative Study of Conservative versus Surgical Treatment Protocols for 77 Patients with Idiopathic Granulomatous Mastitis. Breast J. 2015;21(4):363–9. doi: 10.1111/tbj.12415 [ DOI ] [ PubMed ] [ Google Scholar ] 29. Shojaee L, Rahmani N, Moradi S, Motamedi A, Godazandeh G. Idiopathic granulomatous mastitis: challenges of treatment in iranian women. BMC Surg. 2021;21(1):206. doi: 10.1186/s12893-021-01210-6 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 30. Akcan A, Oz AB, Dogan S, Akgün H, Akyüz M, Ok E, et al. Idiopathic Granulomatous Mastitis: Comparison of Wide Local Excision with or without Corticosteroid Therapy. Breast Care (Basel). 2014;9(2):111–5. doi: 10.1159/000360926 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 31. Shin YD, Park SS, Song YJ, Son S-M, Choi YJ. Is surgical excision necessary for the treatment of Granulomatous lobular mastitis?. BMC Womens Health. 2017;17(1):49. doi: 10.1186/s12905-017-0412-0 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 32. Montazer M, Dadashzadeh M, Moosavi Toomatari SE. Comparison of the Outcome of Low Dose and High-Dose Corticosteroid in the Treatment of Idiopathic Granulomatous Mastitis. Asian Pac J Cancer Prev. 2020;21(4):993–6. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 33. Karanlik H, Ozgur I, Simsek S, Fathalizadeh A, Tukenmez M, Sahin D, et al. Can Steroids plus Surgery Become a First-Line Treatment of Idiopathic Granulomatous Mastitis?. Breast Care (Basel). 2014;9(5):338–42. doi: 10.1159/000366437 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] PLoS One. doi: 10.1371/journal.pone.0341901.r001 Decision Letter 0 Siddharth Gosavi Siddharth Gosavi Academic Editor Find articles by Siddharth Gosavi Author information Copyright and License information Roles Siddharth Gosavi : Academic Editor © 2026 Siddharth Gosavi This is an open access article distributed under the terms of the Creative Commons Attribution License , which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited., which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. PMC Copyright notice 12 Dec 2025 Dear Dr. Luo, Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process. Please submit your revised manuscript by Jan 26 2026 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at [email protected] . When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file. A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'. A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'. An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'. If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter. If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: https://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols . Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at . Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols .. We look forward to receiving your revised manuscript. Kind regards, Siddharth Gosavi, MBBS, MD Internal Medicine,DNB Internal Medicine Academic Editor PLOS One Journal requirements: When submitting your revision, we need you to address these additional requirements. 1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf 2. In the online submission form you indicate that your data is not available for proprietary reasons and have provided a contact point for accessing this data. Please note that your current contact point is a co-author on this manuscript. According to our Data Policy, the contact point must not be an author on the manuscript and must be an institutional contact, ideally not an individual. Please revise your data statement to a non-author institutional point of contact, such as a data access or ethics committee, and send this to us via return email. Please also include contact information for the third party organization, and please include the full citation of where the data can be found. If the reviewer comments include a recommendation to cite specific previously published works, please review and evaluate these publications to determine whether they are relevant and should be cited. There is no requirement to cite these works unless the editor has indicated otherwise. Additional Editor Comments: This large-scale retrospective cohort study, which examined 614 patients diagnosed with GLM between 2017 and 2024, evaluated the effect of prednisolone on relapse in terms of treatment start time, duration, and starting dose. Analysis showed that oral prednisolone use was an independent risk factor for relapse in GLM (HR 2.534). The one-year relapse rate was significantly higher in patients given prednisolone (29.4% vs. 14.6%). However, subgroup analyses showed that starting treatment within 6 weeks of symptom onset and using a dose of ≤0.4 mg/kg/day significantly reduced relapse rates (1-year relapse 6.2%). This rate was similar to the group not receiving prednisolone. In contrast, early treatment initiation (0–6 weeks) was associated with the highest relapse rate (33.5%). In conclusion, although prednisolone generally increases the risk of relapse in GLM, delayed start and low-dose protocols appear to be the most appropriate treatment strategy to reduce this risk. İntro: The introduction provides a well-positioned, logical scientific basis. Clarify the purpose in the final section, aligning it with the study design. MM&Results Major Comments The manuscript combines retrospective data (2017–2022) with prospectively collected data (from 2022 onward). This hybrid design introduces potential inconsistencies in data completeness and diagnostic/treatment protocols across time periods. The authors should clarify: *how data collection was standardized across the retrospective and prospective phases, *whether the two periods differ significantly in baseline characteristics or treatment strategies, *whether a sensitivity analysis excluding early retrospective years was performed. Prednisone administration was not randomized and may reflect clinician judgment based on disease severity. This introduces substantial bias: *severe cases may have been preferentially treated earlier or with higher doses, *the higher recurrence rate in the prednisone group may therefore reflect baseline severity rather than treatment effect. Different analyses include different patient counts (e.g., 335 for timing analysis, 341 for duration, 337 for dose). The mechanism of missing data (MCAR/MAR/MNAR) is not described, and there is no imputation strategy. Given that missingness may introduce bias, the authors should: *report missing data percentages for each variable, *describe the missing data mechanism, *and consider multiple imputation if appropriate. The manuscript uses X-tile to identify cutoffs for treatment timing, dose, and duration. This method increases the risk of overfitting and inflated type I error, especially without an external validation cohort. The authors should clarify: *whether correction for multiple testing was applied, *whether cutoffs were validated using bootstrap or internal cross-validation, *and how robust these thresholds are across sensitivity analyses. Patients received diverse additional therapies, including antimicrobials, anti-tuberculosis agents, drainage procedures, and bromocriptine. Many of these are likely to influence recurrence risk. The multivariable Cox model includes some of these factors, but: *their distribution between exposed and unexposed groups is not reported, *no stratified or interaction analyses are provided, *the possibility of treatment clustering (e.g., prednisone + drainage) is not explored. *A more granular adjustment or subgroup analysis is warranted. “Return of symptoms after clinical improvement or cure” is somewhat subjective and operator-dependent. Since recurrence is the primary endpoint, the authors should: *specify objective criteria (e.g., imaging, physical examination, laboratory data), *clarify whether recurrence was assessed by blinded observers, *and report inter-observer variability if applicable. Some subgroups used in the final survival analyses appear very small due to filtering (e.g., n=1 or n=2 cases added). This reduces reliability of hazard estimates. The authors should: *report subgroup sample sizes clearly, *refrain from overly granular stratification when numbers are insufficient, *discuss the limitations in the power of subgroup comparisons. Minor Comments The dosing description (“initiated at a high dose and tapered by 5 mg every 3–7 days”) lacks: *the definition of “high dose,” *typical taper duration, *criteria for tapering speed. *A standardized protocol description is needed for reproducibility. Figure 1 outlines the workflow but does not provide a CONSORT-style patient flow diagram. A flowchart of inclusions/exclusions and analysis subsets would greatly improve transparency. The manuscript lacks a comparison of baseline characteristics between: *prednisone vs. non-prednisone groups, *short- vs. long-interval groups, *low- vs. high-dose groups. Rationale for Dividing Dosage by Both Absolute and Weight-Adjusted Methods: Both approaches are analyzed, but the rationale and clinical relevance should be explained more succinctly. Language: The Methods section is thorough but dense. Some subsections may benefit from condensation for readability. In summary; This manuscript is strengthened by a large sample size and a comprehensive multidimensional analysis of prednisone therapy. However, concerns regarding confounding, missing data, exposure classification, and multiple threshold testing must be addressed before the conclusions can be fully supported. Clarification of the methodology and additional analyses as suggested above would considerably improve the manuscript’s rigor and impact. Discussion: The discussion is generally aligned with the study's three main objectives: treatment timing, dose, and duration; this is good. However, although the reasons for choosing these three dimensions—the clinical relevance of the research gap—were strongly emphasized in the introduction, the discussion occasionally becomes disjointed. Some lengthy explanations (e.g., pathophysiology, steroid rebound mechanism) distract from the study's main question. The comparison with the literature is comprehensive; many studies are compared. However, the comparisons are presented in very long passages, somewhat obscuring the clinical implications. In some sections, citations are listed consecutively, but the question "What does this study add to the literature?" is not clearly stated. The proposed treatment strategy at the end of the discussion is a strong point; however, more targeted conclusions, such as for whom it should be applied and for which clinical phenotypes, would have added clinical value. Limitations are accurately listed. However, the heterogeneity of surgical interventions and the impact of nonsteroidal therapies, particularly as a critical limitation, should have been integrated into the discussion earlier. IGM is a disease with many unknown issues, especially regarding its treatment. Studies including a large number of patients, like this one, could be very helpful in clarifying the concern of using steroids and their adverse effects in IGM. The three dimensions considered for grouping treatment features and the cut-off point measurements are interesting, providing a well-designed analysis protocol. However, there are many concerns that need to be addressed before publication could be considered for this manuscript: 1. ABSTRACT: needs some English language editing. Also, the main concerns that will be approached below should be mentioned in brief in the abstract. INTRODUCTION: 2. All through the introduction, the past tense is wrongly used for present data on GLM. 3. Lines 44 to 47: The authors mention that IGM incidence has rapidly increased during the present decade. For this, they cite a paper from 2019, which is a clinical study on the surgical treatment of GLM. That papers cite, for stating the incidence conditions, a paper from 2009, which reports a series of 54 cases of GLM. None of these investigates the incidence of GLM, and the year of publication of the latter is too old to talk about the recent decade. Therefore, this sentence is not supported by the present evidence. To update the issue, please know that the only papers published up to now (as far as I know) with an acceptable assessment of GLM incidence are those of the US CDC [Goldman M, et al. Idiopathic Granulomatous Mastitis in Hispanic Women--Indiana, 2006-2008. MMWR: Morbidity & Mortality Weekly Report. 2009 Dec 4;58(47)], which only includes Indiana (the city with the highest Hispanic population in the US) and cannot be applicated globally, and that of Zou et al [Zou J, et al. Clinical trends in granulomatous mastitis incidence, prevalence, and treatment: A retrospective study highlighting ethnic differences in care. Journal of Surgical Research. 2024 Oct 1;302:732-8.] which consists of a population-based study in the USA. The results of the latter are very valuable, but can only be applied to the US population, and they do not assess the recent incidence trends. So, the data in these lines of the manuscript lack evidence, unless the authors provide a correct reference. 4. Lines 55-57: ‘Although topical steroid therapy was reported as a useful treatment modality [6,11], its invasive nature limited its widespread clinical adoption.’ Topical steroids are the least invasive treatment, how do authors mention that the reason for its limited use is its invasive nature? 5. Lines 52-67: While using many references about it, the authors forgot to mention the recently adopted local (intralesional) use of steroids. May be, by topical use and its invasiveness, they meant local injections?? 6. Introduction overall: The introduction is not well-written. Authors could have a better approach to the present therapeutic approaches and the existing challenges in GLM treatment. Also, despite the large amount of literature available, some references are selected inappropriately, like the use of references 1 (comparing local and oral steroids in GLM treatment) and 2 (review of GLM pathology) for the first sentence (‘ Granulomatous lobular mastitis was a rare, benign, and chronic inflammatory disease of the breast with an unknown etiology’). METHODS 7. Lines 80-83: Before mentioning the number of excluded patients, the inclusion and exclusion criteria should be stated clearly. 8. Line 84: 1) What was the mode of diagnosis of GLM? What were the diagnostic criteria?2) Did the included patients used only prednisolone as their treatment modality? How did the authors get sure about the concomitant use of NSAIDs, or any combination treatment? 9. Please explain when were the lab tests done? As the test results were related to disease prognosis, the timing of the tests is very important. Usually, all GLM patients do not attend at the first disease presentation. Therefore, tests are taken at different disease stages. How was the condition of lab test timing in your database? If the mentioned concerns existed in your study (as expected), please mention this limitation in the discussion. 10. Fig.1: Although well-organized and very useful, I suggest that sentences are shortened. Usually, short sentences or expressions are used in these types of workflows. E.g., instead of ‘614 pts with GLM were ultimately included in the analysis’, a shortened version as ’Finally included: 614 patients’ would be more suitable for the figure. This should be done in other parts of the figure also. Importantly, the ’It was hypothesized that …’ is not appropriate for the figure. It could be replaced by something like ’Assessment of prednisolone effects’. 11. Line 144: How do authors explain the effect of the affected side on prognosis? 12. Tab 1: For prolactin, please provide the normal range. Also, define hyperlipidemia (which types of lipids? The ranges?). Was bromocriptine used for hyperprolactinemia? These should be described shortly in the footnote to make the table understandable and useful. 13. Tab. 1: For the age at diagnosis, which carried a worse prognosis, younger or elder age? 14. Lines 152-154: These descriptions should come in the Methods, not the results. Also, please explain what were the non-prednisone treatments. The overall treatment strategies are described in lines 97-100; however, we need to know, for example, if you had patients who only received antibiotics? Or only anti-tuberculosis agents? Or if bromocriptine was used alone, and regardless of the prolactin status? Regarding the higher recurrence in the prednisone group (as mentioned in lines 154-156 and Fig 2), it is crucial to know about the comparison group details. DISCUSSION 15. Lines 283-285: The point is that the two groups: prednisone and non=prednisone are not compared regarding disease severity, so the results might just mirror the disease severity. Please explain if that is not the point, and otherwisw discuss the point as such and add it in the limitations. 16. Lines 309-327: These are hypothetical reasons for the higher recurrence in patients with early-onset treatment, and they do not follow a very sound rational. Patients who attended later probably had a less severe disease, and those that attended earlier had a more severe disease. So, treatment was begun sooner, but recurrence was higher. Although this is not certain, but it should be mentioned as a probable reason for these results, and recognized as a main confounding factor in the discussion and limitations. 17. Lines 328-331: Treatment duration depends on the disease response to treatment, so medicines cannot be cut before obtaining an acceptable response. Do the authors mean to discontinue steroids while the disease is improving but we are still waiting for recovery? The reason for the lower recurrence with a shorter treatment duration might simply be because of less severe disease. Authors should consider this important point in the discussion and limitations. 18. Lines 332 to 358: This part should be shortened and simplified with an accurate wording. Authors should pay attention to the tense of the verbs and the use of ‘this’ and ‘that’ for theirs vs. other studies. 19. Lines 367-369: Lower doses are most probably due to less severe disease, which explains the lower recurrence. As the clinical signs are not compared between the two groups, we cannot know whether the low and high dose groups were similar in this regard, and they were most probably not. This also applies to the weight-adjusted dose. This key point should be considered and alluded to in the discussion and limitations. 20. Apparently, the results and discussion indicate that the non-prednisone treatments, the lower doses of prednisone, the later-onset treatment initiation, and the shorter duration of prednisolone use is either better that or not different with the comparison group better. Wouldn’t all these reflect that prednisolone is better not used? 21. Overall, as disease severity has not been considered in the comparison between groups, results should be taken cautiously. The very positive sentences that present results of this study as doubtless data need to be re-written to reflect the uncertainty of the mentioned points. The limitations also should be revised accordingly. [Note: HTML markup is below. Please do not edit.] Reviewers' comments: Reviewer's Responses to Questions Comments to the Author 1. Is the manuscript technically sound, and do the data support the conclusions? Reviewer #1: Yes Reviewer #2: Yes Reviewer #3: Partly ********** 2. Has the statistical analysis been performed appropriately and rigorously? -->?> Reviewer #1: Yes Reviewer #2: Yes Reviewer #3: Yes ********** 3. Have the authors made all data underlying the findings in their manuscript fully available??> The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.--> Reviewer #1: Yes Reviewer #2: No Reviewer #3: Yes ********** 4. Is the manuscript presented in an intelligible fashion and written in standard English??> Reviewer #1: Yes Reviewer #2: Yes Reviewer #3: Yes ********** Reviewer #1: The study is well designed and it will help to physicians to use oral steroid therapy for the patients with IGM. In fact I use intralesional steroid therappy in the treatment of IGM. However this study has investigated various therapies for the patients and explained the dosage and the timing of initiation of oral steroid therapy. So, it is available for publication. Reviewer #2: This large-scale retrospective cohort study, which examined 614 patients diagnosed with GLM between 2017 and 2024, evaluated the effect of prednisolone on relapse in terms of treatment start time, duration, and starting dose. Analysis showed that oral prednisolone use was an independent risk factor for relapse in GLM (HR 2.534). The one-year relapse rate was significantly higher in patients given prednisolone (29.4% vs. 14.6%). However, subgroup analyses showed that starting treatment within 6 weeks of symptom onset and using a dose of ≤0.4 mg/kg/day significantly reduced relapse rates (1-year relapse 6.2%). This rate was similar to the group not receiving prednisolone. In contrast, early treatment initiation (0–6 weeks) was associated with the highest relapse rate (33.5%). In conclusion, although prednisolone generally increases the risk of relapse in GLM, delayed start and low-dose protocols appear to be the most appropriate treatment strategy to reduce this risk. İntro: The introduction provides a well-positioned, logical scientific basis. Clarify the purpose in the final section, aligning it with the study design. MM&Results Major Comments The manuscript combines retrospective data (2017–2022) with prospectively collected data (from 2022 onward). This hybrid design introduces potential inconsistencies in data completeness and diagnostic/treatment protocols across time periods. The authors should clarify: *how data collection was standardized across the retrospective and prospective phases, *whether the two periods differ significantly in baseline characteristics or treatment strategies, *whether a sensitivity analysis excluding early retrospective years was performed. Prednisone administration was not randomized and may reflect clinician judgment based on disease severity. This introduces substantial bias: *severe cases may have been preferentially treated earlier or with higher doses, *the higher recurrence rate in the prednisone group may therefore reflect baseline severity rather than treatment effect. Different analyses include different patient counts (e.g., 335 for timing analysis, 341 for duration, 337 for dose). The mechanism of missing data (MCAR/MAR/MNAR) is not described, and there is no imputation strategy. Given that missingness may introduce bias, the authors should: *report missing data percentages for each variable, *describe the missing data mechanism, *and consider multiple imputation if appropriate. The manuscript uses X-tile to identify cutoffs for treatment timing, dose, and duration. This method increases the risk of overfitting and inflated type I error, especially without an external validation cohort. The authors should clarify: *whether correction for multiple testing was applied, *whether cutoffs were validated using bootstrap or internal cross-validation, *and how robust these thresholds are across sensitivity analyses. Patients received diverse additional therapies, including antimicrobials, anti-tuberculosis agents, drainage procedures, and bromocriptine. Many of these are likely to influence recurrence risk. The multivariable Cox model includes some of these factors, but: *their distribution between exposed and unexposed groups is not reported, *no stratified or interaction analyses are provided, *the possibility of treatment clustering (e.g., prednisone + drainage) is not explored. *A more granular adjustment or subgroup analysis is warranted. “Return of symptoms after clinical improvement or cure” is somewhat subjective and operator-dependent. Since recurrence is the primary endpoint, the authors should: *specify objective criteria (e.g., imaging, physical examination, laboratory data), *clarify whether recurrence was assessed by blinded observers, *and report inter-observer variability if applicable. Some subgroups used in the final survival analyses appear very small due to filtering (e.g., n=1 or n=2 cases added). This reduces reliability of hazard estimates. The authors should: *report subgroup sample sizes clearly, *refrain from overly granular stratification when numbers are insufficient, *discuss the limitations in the power of subgroup comparisons. Minor Comments The dosing description (“initiated at a high dose and tapered by 5 mg every 3–7 days”) lacks: *the definition of “high dose,” *typical taper duration, *criteria for tapering speed. *A standardized protocol description is needed for reproducibility. Figure 1 outlines the workflow but does not provide a CONSORT-style patient flow diagram. A flowchart of inclusions/exclusions and analysis subsets would greatly improve transparency. The manuscript lacks a comparison of baseline characteristics between: *prednisone vs. non-prednisone groups, *short- vs. long-interval groups, *low- vs. high-dose groups. Rationale for Dividing Dosage by Both Absolute and Weight-Adjusted Methods: Both approaches are analyzed, but the rationale and clinical relevance should be explained more succinctly. Language: The Methods section is thorough but dense. Some subsections may benefit from condensation for readability. In summary; This manuscript is strengthened by a large sample size and a comprehensive multidimensional analysis of prednisone therapy. However, concerns regarding confounding, missing data, exposure classification, and multiple threshold testing must be addressed before the conclusions can be fully supported. Clarification of the methodology and additional analyses as suggested above would considerably improve the manuscript’s rigor and impact. Discussion: The discussion is generally aligned with the study's three main objectives: treatment timing, dose, and duration; this is good. However, although the reasons for choosing these three dimensions—the clinical relevance of the research gap—were strongly emphasized in the introduction, the discussion occasionally becomes disjointed. Some lengthy explanations (e.g., pathophysiology, steroid rebound mechanism) distract from the study's main question. The comparison with the literature is comprehensive; many studies are compared. However, the comparisons are presented in very long passages, somewhat obscuring the clinical implications. In some sections, citations are listed consecutively, but the question "What does this study add to the literature?" is not clearly stated. The proposed treatment strategy at the end of the discussion is a strong point; however, more targeted conclusions, such as for whom it should be applied and for which clinical phenotypes, would have added clinical value. Limitations are accurately listed. However, the heterogeneity of surgical interventions and the impact of nonsteroidal therapies, particularly as a critical limitation, should have been integrated into the discussion earlier. Reviewer #3: Thank you for the opportunity to review this manuscript. IGM is a disease with many unknown issues, especially regarding its treatment. Studies including a large number of patients, like this one, could be very helpful in clarifying the concern of using steroids and their adverse effects in IGM. The three dimensions considered for grouping treatment features and the cut-off point measurements are interesting, providing a well-designed analysis protocol. However, there are many concerns that need to be addressed before publication could be considered for this manuscript: 1. ABSTRACT: needs some English language editing. Also, the main concerns that will be approached below should be mentioned in brief in the abstract. INTRODUCTION: 2. All through the introduction, the past tense is wrongly used for present data on GLM. 3. Lines 44 to 47: The authors mention that IGM incidence has rapidly increased during the present decade. For this, they cite a paper from 2019, which is a clinical study on the surgical treatment of GLM. That papers cite, for stating the incidence conditions, a paper from 2009, which reports a series of 54 cases of GLM. None of these investigates the incidence of GLM, and the year of publication of the latter is too old to talk about the recent decade. Therefore, this sentence is not supported by the present evidence. To update the issue, please know that the only papers published up to now (as far as I know) with an acceptable assessment of GLM incidence are those of the US CDC [Goldman M, et al. Idiopathic Granulomatous Mastitis in Hispanic Women--Indiana, 2006-2008. MMWR: Morbidity & Mortality Weekly Report. 2009 Dec 4;58(47)], which only includes Indiana (the city with the highest Hispanic population in the US) and cannot be applicated globally, and that of Zou et al [Zou J, et al. Clinical trends in granulomatous mastitis incidence, prevalence, and treatment: A retrospective study highlighting ethnic differences in care. Journal of Surgical Research. 2024 Oct 1;302:732-8.] which consists of a population-based study in the USA. The results of the latter are very valuable, but can only be applied to the US population, and they do not assess the recent incidence trends. So, the data in these lines of the manuscript lack evidence, unless the authors provide a correct reference. 4. Lines 55-57: ‘Although topical steroid therapy was reported as a useful treatment modality [6,11], its invasive nature limited its widespread clinical adoption.’ Topical steroids are the least invasive treatment, how do authors mention that the reason for its limited use is its invasive nature? 5. Lines 52-67: While using many references about it, the authors forgot to mention the recently adopted local (intralesional) use of steroids. May be, by topical use and its invasiveness, they meant local injections?? 6. Introduction overall: The introduction is not well-written. Authors could have a better approach to the present therapeutic approaches and the existing challenges in GLM treatment. Also, despite the large amount of literature available, some references are selected inappropriately, like the use of references 1 (comparing local and oral steroids in GLM treatment) and 2 (review of GLM pathology) for the first sentence (‘ Granulomatous lobular mastitis was a rare, benign, and chronic inflammatory disease of the breast with an unknown etiology’). METHODS 7. Lines 80-83: Before mentioning the number of excluded patients, the inclusion and exclusion criteria should be stated clearly. 8. Line 84: 1) What was the mode of diagnosis of GLM? What were the diagnostic criteria?2) Did the included patients used only prednisolone as their treatment modality? How did the authors get sure about the concomitant use of NSAIDs, or any combination treatment? 9. Please explain when were the lab tests done? As the test results were related to disease prognosis, the timing of the tests is very important. Usually, all GLM patients do not attend at the first disease presentation. Therefore, tests are taken at different disease stages. How was the condition of lab test timing in your database? If the mentioned concerns existed in your study (as expected), please mention this limitation in the discussion. 10. Fig.1: Although well-organized and very useful, I suggest that sentences are shortened. Usually, short sentences or expressions are used in these types of workflows. E.g., instead of ‘614 pts with GLM were ultimately included in the analysis’, a shortened version as ’Finally included: 614 patients’ would be more suitable for the figure. This should be done in other parts of the figure also. Importantly, the ’It was hypothesized that …’ is not appropriate for the figure. It could be replaced by something like ’Assessment of prednisolone effects’. 11. Line 144: How do authors explain the effect of the affected side on prognosis? 12. Tab 1: For prolactin, please provide the normal range. Also, define hyperlipidemia (which types of lipids? The ranges?). Was bromocriptine used for hyperprolactinemia? These should be described shortly in the footnote to make the table understandable and useful. 13. Tab. 1: For the age at diagnosis, which carried a worse prognosis, younger or elder age? 14. Lines 152-154: These descriptions should come in the Methods, not the results. Also, please explain what were the non-prednisone treatments. The overall treatment strategies are described in lines 97-100; however, we need to know, for example, if you had patients who only received antibiotics? Or only anti-tuberculosis agents? Or if bromocriptine was used alone, and regardless of the prolactin status? Regarding the higher recurrence in the prednisone group (as mentioned in lines 154-156 and Fig 2), it is crucial to know about the comparison group details. DISCUSSION 15. Lines 283-285: The point is that the two groups: prednisone and non=prednisone are not compared regarding disease severity, so the results might just mirror the disease severity. Please explain if that is not the point, and otherwisw discuss the point as such and add it in the limitations. 16. Lines 309-327: These are hypothetical reasons for the higher recurrence in patients with early-onset treatment, and they do not follow a very sound rational. Patients who attended later probably had a less severe disease, and those that attended earlier had a more severe disease. So, treatment was begun sooner, but recurrence was higher. Although this is not certain, but it should be mentioned as a probable reason for these results, and recognized as a main confounding factor in the discussion and limitations. 17. Lines 328-331: Treatment duration depends on the disease response to treatment, so medicines cannot be cut before obtaining an acceptable response. Do the authors mean to discontinue steroids while the disease is improving but we are still waiting for recovery? The reason for the lower recurrence with a shorter treatment duration might simply be because of less severe disease. Authors should consider this important point in the discussion and limitations. 18. Lines 332 to 358: This part should be shortened and simplified with an accurate wording. Authors should pay attention to the tense of the verbs and the use of ‘this’ and ‘that’ for theirs vs. other studies. 19. Lines 367-369: Lower doses are most probably due to less severe disease, which explains the lower recurrence. As the clinical signs are not compared between the two groups, we cannot know whether the low and high dose groups were similar in this regard, and they were most probably not. This also applies to the weight-adjusted dose. This key point should be considered and alluded to in the discussion and limitations. 20. Apparently, the results and discussion indicate that the non-prednisone treatments, the lower doses of prednisone, the later-onset treatment initiation, and the shorter duration of prednisolone use is either better that or not different with the comparison group better. Wouldn’t all these reflect that prednisolone is better not used? 21. Overall, as disease severity has not been considered in the comparison between groups, results should be taken cautiously. The very positive sentences that present results of this study as doubtless data need to be re-written to reflect the uncertainty of the mentioned points. The limitations also should be revised accordingly. ********** what does this mean? ). If published, this will include your full peer review and any attached files.). If published, this will include your full peer review and any attached files. If you choose “no”, your identity will remain anonymous but your review may still be made public. Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our For information about this choice, including consent withdrawal, please see our Privacy Policy .--> Reviewer #1: No Reviewer #2: Yes: Irmak Durur-SubasiIrmak Durur-Subasi Reviewer #3: Yes: Sadaf AlipourSadaf Alipour ********** [NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.] To ensure your figures meet our technical requirements, please review our figure guidelines: https://journals.plos.org/plosone/s/figures You may also use PLOS’s free figure tool, NAAS, to help you prepare publication quality figures: https://journals.plos.org/plosone/s/figures#loc-tools-for-figure-preparation . NAAS will assess whether your figures meet our technical requirements by comparing each figure against our figure specifications. PLoS One. 2026 Apr 8;21(4):e0341901. doi: 10.1371/journal.pone.0341901.r002 Author response to Decision Letter 1 Article notes Copyright and License information Collection date 2026. PMC Copyright notice 18 Dec 2025 For details, refer to Document Response to Reviewers Attachment Submitted filename: Response to Reviewers.docx pone.0341901.s005.docx (62.2KB, docx) PLoS One. doi: 10.1371/journal.pone.0341901.r003 Decision Letter 1 Siddharth Gosavi Siddharth Gosavi Academic Editor Find articles by Siddharth Gosavi Author information Copyright and License information Roles Siddharth Gosavi : Academic Editor © 2026 Siddharth Gosavi This is an open access article distributed under the terms of the Creative Commons Attribution License , which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited., which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. PMC Copyright notice 14 Jan 2026 The paradoxical effect of oral prednisone in granulomatous lobular mastitis: an observational cohort study. PONE-D-25-57872R1 Dear Dr. Luo, We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements. Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication. An invoice will be generated when your article is formally accepted. Please note, if your institution has a publishing partnership with PLOS and your article meets the relevant criteria, all or part of your publication costs will be covered. Please make sure your user information is up-to-date by logging into Editorial Manager at Editorial Manager® and clicking the ‘Update My Information' link at the top of the page. For questions related to billing, please contact  and clicking the ‘Update My Information' link at the top of the page. For questions related to billing, please contact billing support .. If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact [email protected]. Kind regards, Siddharth Gosavi, MBBS, MD Internal Medicine,DNB Internal Medicine Academic Editor PLOS One Additional Editor Comments (optional): Satisfied with the response. Reviewers' comments: PLoS One. doi: 10.1371/journal.pone.0341901.r004 Acceptance letter Siddharth Gosavi Siddharth Gosavi Academic Editor Find articles by Siddharth Gosavi Author information Copyright and License information Roles Siddharth Gosavi : Academic Editor © 2026 Siddharth Gosavi This is an open access article distributed under the terms of the Creative Commons Attribution License , which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited., which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. PMC Copyright notice PONE-D-25-57872R1 PLOS One Dear Dr. Luo, I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS One. Congratulations! Your manuscript is now being handed over to our production team. At this stage, our production department will prepare your paper for publication. This includes ensuring the following: * All references, tables, and figures are properly cited * All relevant supporting information is included in the manuscript submission, * There are no issues that prevent the paper from being properly typeset You will receive further instructions from the production team, including instructions on how to review your proof when it is ready. Please keep in mind that we are working through a large volume of accepted articles, so please give us a few days to review your paper and let you know the next and final steps. Lastly, if your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact [email protected]. You will receive an invoice from PLOS for your publication fee after your manuscript has reached the completed accept phase. If you receive an email requesting payment before acceptance or for any other service, this may be a phishing scheme. Learn how to identify phishing emails and protect your accounts at https://explore.plos.org/phishing . If we can help with anything else, please email us at [email protected]. Thank you for submitting your work to PLOS ONE and supporting open access. Kind regards, PLOS ONE Editorial Office Staff on behalf of Dr. Siddharth Gosavi Academic Editor PLOS One Associated Data This section collects any data citations, data availability statements, or supplementary materials included in this article. Supplementary Materials S1 Table. Univariable and multivariable Cox regression analysis restricted to the prospective cohort (n = 374). (DOCX) pone.0341901.s001.docx (27.2KB, docx) S2 Table. Baseline characteristics of patients before and after propensity score matching. (DOCX) pone.0341901.s002.docx (28KB, docx) S1 Checklist. PLOSOne clinical studies checklist. (DOCX) pone.0341901.s003.docx (34.1KB, docx) Attachment Submitted filename: Response to Reviewers.docx pone.0341901.s005.docx (62.2KB, docx) Data Availability Statement The datasets generated and analyzed in this study are not publicly available as they belong to a proprietary clinical database. Access can be granted for research purposes upon formal request. Interested researchers should contact the institutional data steward, Director Hu (Department of Breast and Thyroid Surgery), at [email protected] . Applicants were asked to specify the intended use of the data and to commit to compliance with the terms of use. Any data sharing was contingent upon final approval from the relevant institutions and the signing of a legally binding data use agreement. Articles from PLOS One are provided here courtesy of PLOS ACTIONS View on publisher site PDF (1.8 MB) Cite Collections Permalink PERMALINK Copy RESOURCES Similar articles Cited by other articles Links to NCBI Databases Cite Copy Download .nbib .nbib Format: AMA APA MLA NLM Add to Collections Create a new collection Add to an existing collection Name your collection * Choose a collection Unable to load your collection due to an error Please try again Add Cancel Follow NCBI NCBI on X (formerly known as Twitter) NCBI on Facebook NCBI on LinkedIn NCBI on GitHub NCBI RSS feed Connect with NLM NLM on X (formerly known as Twitter) NLM on Facebook NLM on YouTube National Library of Medicine 8600 Rockville Pike Bethesda, MD 20894 Web Policies FOIA HHS Vulnerability Disclosure Help Accessibility Careers NLM NIH HHS USA.gov Back to Top