Mastitis is a common inflammatory disease of the breast, which could occur in all populations.[1–4] Depending on the association with breastfeeding, mastitis can be classified into puerperal mastitis (PM) and non-puerperal mastitis (NPM). These 2 types have different etiologies, bacteriologies, presentations, and treatments. According to previous studies, incidence of the PM among breastfeeding women varies enormously from 2.6% to 33%. The 2 basic causes of PM are milk stasis and infection. Lactational abscesses were significantly more likely to be caused by Staphylococcus aureus.[6,7] Other documented microorganisms are Bacteroids spp, Escherichia coli, or other gram-negative bacteria, group A and group B hemolytic streptococci, Peptostreptococcus spp, and Mycobacterium tuberculosis (rare). Moreover, Candida albicans was also shown to be detected in human milk, but not all of these women developed an infection.[8,9]
Compared with the PM, NPM is a rare inflammatory breast disease. NPM is a benign inflammatory disease that commonly recurs, and is always with formation of a mammary duct fistula.[10,11–13] Appropriate diagnosis and treatment for NPM is complicated and has been under controversy for some time. According to the histopathologic types, NPM can be roughly divided into 2 types: periductal mastitis (PDM)/mammary ductectasia (MDE) and idiopathic granulomatous mastitis (IGM). PDM affects non-lactating women between the ages of 19 to 48 years and is characterized by periareolar inflammation with or without a mass, a periareolar abscess, or a mammary duct fistula. On account of the difficulty in the differential diagnosis of PDM and MDE clinically and pathologically, the PDM and MDE patients were merged into a single PDM/MDE group in our study. IGM is a rare inflammatory disease of the breast, affecting mainly women of childbearing age. Because it shares similarity with breast carcinoma, the diagnosis of IGM remains challenging in clinical practice. It has been estimated that between 0.44% and 1.6% of breast biopsy specimens represent IGM based on accepted pathological criteria.
The aim of this retrospective study was to analyze the clinicopathologic characteristics and bacteriological etiology of non-lactational mastitis, as well as to seek an effective diagnosis and management strategy for NPM.
Subjects and methods
In this retrospective study, the clinical and histopathologic data of consecutive patients with NPM including PDM/MDE and IGM patients in the Breast Tumor Centre of Sun Yat-Sen Memorial Hospital, China between January 2005 and May 2016, were collected.
In this study, inclusion criteria are as follows: (1) patients with histologically confirmed mastitis including PDM/MDE and IGM, (2) first onset and recurrence of NPM, and (3) a population in Southern China. Exclusion criteria include (1) being pregnant or breastfeeding, (2) specific granulomatous diseases, (3) male patients with mastitis, (4) complicated with acute or chronic infection in other sites, and (5) history of breast cancer or other human cancers.
The study was approved by the institutional review board of the Breast Tumor Centre of Sun Yat-Sen Memorial Hospital, China. The following data were included in the study: clinical features, laboratory studies, details of imaging, and microbiological studies. Further details regarding treatment and follow-up were also collected.
In ultrasonography, bilateral breast ultrasonic scanning was performed to detect possible lesions. The data including location, maximum diameter, echo, shape, margin, and vascularity were recorded. In mammography, low energy x-ray was used for breast imaging. The date including location, maximum diameter lesion characteristics were recorded. All breast lesions were graded by 2 experienced ultrasound doctors and radiologists according to grading criteria of the Breast Imaging Reporting and Data System (BI-RADS).
Moreover, serology tests including complete blood count, erythrocyte sedimentation rate (ESR), high-sensitivity C-reactive protein (hsCRP), electrolytes, and liver and renal functional tests were performed in all patients. The purified protein derivative test and serum anti-tuberculosis drug levels tested by the enzyme-linked immunosorbent assay were performed to help diagnosis.
The definitive pathology examination was processed as the essential criterion in diagnosis of NPM. All ablated portions of the mammary tissue from core-needle biopsies and open surgical biopsies were widely resected and placed in 10% formalin and stored at 4°C until sectioning. All aspirates and tissue samples were examined with hematoxylin–eosin staining procedure as well as the special stains for tuberculosis and fungal infection. Histologically, IGM is characterized by non-necrotizing granuloma formation with a localized infiltrate of multi-nucleated giant cells, plasma cells, epithelioid histiocytes, lymphocytes, and fibrous hyperplasia. As for PDM/MDE, it is characterized by periductal inflammation around ectatic ducts showing some disruption of the surrounding elastic lamina with the lymphocytes, plasma cells, and macrophages infiltrated.
Microbiology study was performed in the patients with breast abscesses or inflammatory breast tissues. The pus taken from breast abscesses or inflammatory breast tissues obtained from biopsies were cultured on blood agar, chocolate agar, fastidious anaerobic thioglycolate broth for bacteria, mycobacteria, and anaerobes. Biochemical identification was possible using a wide battery of methods including API Coryne V2.0, API Strep, API NH, API NE, Vitek2 microbiology analyzer (bioMérieux, Marcy-l’Etoile, France), matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (bioMérieux) and 16S rRNA gene amplification and sequencing. Antibiotic susceptibility was determined by the disk diffusion method and E-test (bioMérieux) on Mueller–Hinton agar supplemented with 5% horse blood. Susceptibility to antibiotics was interpreted following recommended criteria by the Clinical and Laboratory Standards Institute (CLSI; formerly, National Committee for Clinical Laboratory Standards) guidelines. Meanwhile, gram staining and acid-fast staining were performed in all samples to find whether there is evidence for the presence of bacteria under the oil microscope.
The patients were categorized into the PDM group and the IGM group for comparison. SPSS, version 19, for windows (IBM, Armonk, NY) was employed for statistical analysis. Chi-square test was employed for testing statistical significance between 2 discrete variables. Measurement data were in mean ± standard deviation. Results were considered significant with a P < .05. All susceptibility data were analyzed using WHONET software, version 5.6.
Demographic and baseline characteristics of the patients with PDM/MDE or IGM
During the study period, 437 patients were admitted because of breast inflammatory disease. According to the inclusion and exclusion criteria, 364 patients with NPM including 117 with PDM/MDE and 247 with IGM were included in this study. Seventy-three patients were excluded due to the diagnosis of lactational mastitis, lack of details, or other reasons for failure to meet the inclusion criteria.
In the PDM/MDE and IGM groups, the median age of the patients was 38.3 (range 15–76) and 35.5 (range 13–70), respectively. Moreover, 46.2% (54/117) of patients in the PDM/MDE group were above 40 years old, while it only account for 27.5% (68/247) in the IGM group. The detailed demographic and personal features of the 2 groups are noted in Table 1. There were no significant differences in the marriage status, breastfeeding history, reproductive history, or age of menarche between the 2 groups. However, there were significant differences in terms of menstruation period and menstruation history. Patients in the PDM/MDE group have a longer medium menstruation period and higher irregular menstruation history rate. No patient in either group has the long-term smoking or alcohol consuming history.
Clinicopathological characteristics of the patients with PDM/MDE or IGM
The most common presenting symptoms were breast lump (83.8% vs 90.3%) and pain (51.3% vs 49.0%) in the PDM/MDE and IGM groups. The left breast was more frequently affected than the right one in all patients. The ratio of clinical suspicion of malignancy was significantly higher in IGM group compared with the PDM/MDE group (Table 2).
Breast ultrasound was the first step radiological investigation in all 117 patients in the PDM/MDE group and 247 patients in the IGM group, whereas mammography could only be performed in 42 patients in the PDE/MDE group and 46 in the IGM group due to the limitation of age or severe breast pain.
The ratio of abscess and fistula was significantly higher in the PDM/MDE group compared with the IGM group (38.5% vs 22.7%, P = .002; 22.2% vs 11.3%, P = .006, respectively). Sonographic examination of the breast revealed the sites of breast lesions were significantly different between the 2 groups. The retroareolar quadrant (34.2%) was the most common site in the PDM/MDE group, while the upper-outerquadrant (27.1%) was the most common site in the IGM. However, there were no significant differences between the findings of mammography between the 2 groups (Table 3).
As for the laboratory parameters, leukocyte count, neutrophil percentage and the rise rate of hsCRP were significantly higher in the PDM/MDE group compared with the IGM group (7.3 ± 2.6 vs 8.0 ± 3.0, P = .027; 62.7 ± 10.1 vs 66.0 ± 10.4, P = .005; 17.0% vs 31.7%, P = .010, respectively), while the level of hemoglobin, lymphocyte percentage, platelet count, and ESR were not statistically significant between the 2 group. Moreover, there were no significant differences in serum levels of tumor markers, including carcinoembryonic antigen (CEA), cancer antigen 125 (CA125), CA153, and cytokeratin 19 fragment (CYFRA21-1) between the 2 groups. The prolactin (PRL) was measured in 75 patients in the PDM/MDE group and 113 in the IGM group. The rise rate of PRL level was significantly higher in the PDM/MDE group compared with the IGM group (Table 4).
Bacteriological etiology of the patients with PDM/MDE or IGM
Microbiology study of 127 patients who were diagnosed with NPM, including 34 with PDM/MDE and 93 with IGM, was further investigated. The Gram and acid-fast stains were performed in all samples from the abscesses or inflammatory breast tissues. Two specimens were positive for Gram stain in the PDM/MDE group, while only 1 specimen was positive in the IGM group. Acid-fast bacteria-Ziehl-Neelsen Stain showed negative for Mycobacterium tuberculosis. The culture results showed bacterial growth in 20 out of 34 cases (58.8%) in the PDM/MDE group and 41 out of 93 (44.0%) in the IGM group. There was a significant difference in the culture positivity (P = .000).
There were 19 patients of infection with single microorganism (95.0%) and 1 case of mixed infection (5.0%) in the PDM/MDE group. A total of 10 organisms were isolated from these 20 cases. The main pathogen among patients with single infection was Staphylococcus (55.0%, 11/20), in which mostly were coagulase-negative staphylococcus (MRCNS) (30%, 6/20), followed by Staphylococcus aureus (S aureus) (25.0%, 5/20) and Corynebacterium kroppenstedtii (25.0%, 5/20). In the IGM group, there were 38 cases of infection with single microorganism (92.7%) and 3 cases of mixed infection (7.3%), a total of 11 organisms were isolated from these 41 cases. The most common organisms identified were Corynebacteriumin (58.5%, 24/41), followed by S aureus in (22.0%, 9/41), and methicillin-resistant coagulase-negative staphylococci (MRCNS) (19.5%, 8/41) (Table 5). C kroppenstedtii, the most commonly microorganisms, isolated from patients in the IGM group was higher compared with the PDM/MDE group (25.0% vs 58.5%, P < .05).
Among the 29 Staphylococcus isolates, only 5 (17.1%) were sensitive to penicillin and 15 (51.4%) to oxacillin. Rates of sensitivity to erythromycin, clindamycin and tetracycline were 51.4%, 60.0% and 68.6%, respectively, while were >80% for ciprofloxacin, levofloxacin, trimethoprim/sulfamethoxazole, gentamicin, and rifampicin. All isolates were susceptible to linezolid, quinupristin/dalfopristin, and vancomycin (Fig. 1). The level of sensitivity to oxacillin was significantly higher in the PDM/MDE group compared with the IGM group (83.3% vs 36.0%, P < .05). Staphylococcus strains isolated from the PDM/MDE group had higher sensitivity than the IGM group to the non-β-lactams antibiotics, including erythromycin (75.0% vs 44.0%, P < .05), clindamycin (75.0% vs 48.0%, P < .05), ciprofloxacin (100% vs 80.0%, P < .05), levofloxacin (100% vs 84.0%, P < .05), teicoplanin (100% vs 72.0%, P < .05), rifampicin (100% vs 72.0%, P < .05), and trimethoprim/sulfamethoxazole (100% vs 72.0%, P < .05). However, tetracycline sensitivity was less prevalent among Staphylococcus isolates from the PDM/MDE group compared with the IGM group (58.3% vs 76.0%, P < .05), whereas there were no detectable differences in the remaining antimicrobials (Fig. 2).
Among the 29 Corynebacterium isolates, rates of sensitivity to erythromycin, clindamycin, and penicillin were 5.0%, 5.0%, and 50.0%, respectively, while were >75% for tetracycline, rifampicin, vancomycin, quinupristin/dalfopristin, and meropenem. All isolates were susceptible to linezolid, ceftriaxone, cefotaxime, cefepime, and ciprofloxacin (Fig. 3).
Due to the frequent mimicking breast carcinoma and chronic recurrence, NPM pose a diagnostic and therapeutic dilemma. Furthermore, NPM is not a single specific histopathological entity, but includes a wide spectrum of different and complex disorders. A variety of entities have been reported to classify the NPM. PDM, MDE, and IGM are the most common subtypes. Often, patients diagnosed as PDM or MDE have similar baseline characteristics and same therapies. PDM was considered sequel of duct ectasia and these probably reflect different stages in 1 disease process. Since previous studies on NPM were always case reports or existed small sample size problem, our study, which is concentrated on clinicopathologic characteristics and bacteriological etiology, may be novel and valuable in providing direct evidence for making wise decisions.
IGM usually affects females under 40 years old,[15,16] which was reflected in our study (the mean age of our patients was 35.5 years). The PDM/MDE is more prevalent in women 30 to 70 years old, which was reflected in our study (the mean age of our patients was 38.3 years). We further found that the median age in PDM/MDE patients is significantly higher compared with IGM patients, which was consistent with previous studies.[18,19] Compared with IGM patients, longer menstruation period and higher ratio of menstruation history were implicated in PDM/MDE patients. Long period of menstruation and hormone disorder contribute to the increased risk of physiologic and pathologic mammary duct expansion. A positive relationship between duct ectasia and tobacco was first described by the Manchester Group and was confirmed by several studies. However, none of the patients were smokers in the present study and the smoking may not be associated with NPM in Chinese women. This is similar to the result of an Indian study.
Nipple discharge and inverted nipple were more spontaneous in women with PDE/MDE compared with IGM, and the difference was statistically significant. Also, the ratio of abscess and fistula in PDM/MDE patients were significant higher than IGM patients. It is suggested that combination of clinical symptoms and ultrasonographic findings may provide a potentially effective approach to differential diagnosis of NPM subtypes. As with the breast cancer, breast lump was the most common presenting symptom. NPM is frequently mistaken for breast carcinoma and pathological confirmation is required for its diagnosis. The laboratory indexes were also measured in NPM patients. The inflammatory biomarkers including leukocyte count, neutrophil percentage and hsCRP in the PDM/MDE group were significantly higher compared with the IGM group. This suggests that the inflammatory process is more severe in PDM/MDE patients than IGM patients.
PRL is a neuroendocrine polypeptide hormone, which stimulates DNA synthesis, epithelial cell proliferation, and milk production in the breast physiologically. The influence of elevated plasma PRL levels on the risk of human breast cancers have been studied by several studies.[24,25] However, the relationship between elevated serum PRL level and the risk of NPM is rarely studied.[26,27] PRL has been shown to promote an inflammatory response in bovine mammary epithelial cells via nuclear factor kappa B activation, and suggest a role for PRL in the pathogenesis of chronic mastitis. However, the regulating mechanism of PRL in human chronic mastitis remains unclear. Interestingly, we found that the rise rate of PRL level in the PDM/MDE group was significantly higher than the IGM group. The hyperprolactinemia may induce and accelerate the dilated lactiferous ducts.
Although bacteria, particularly anaerobic organisms, appear to play some roles in the initiation and progression, the hypothesis that the genesis of NPM may be an infectious inflammatory process has not been clearly documented. Rupture of ectatic ducts or cysts can arise from chemical inflammation in the breast tissue, which can be superimposed by a secondary bacterial infection.[28,29] We were able to obtain a medium frequency of positive cultures in both groups of this study, representing 58.8% and 44.0% of patients with PDM/MDE and IGM, respectively. Bundred et al isolated organisms from 62% of PDM/MDE patients and Ramalingam et al isolated the pathogens in nearly 46% of PDM/MDE patients. We were able to obtain a medium frequency of positive cultures representing 58.8% in the PDM/MDE group in this study. These findings may support the role of infection in the pathogenesis of PDM/MDE. Like the predominance of Staphylococcus sp in aerobic cultures were observed by previous studies,[22,31] the most commonly isolated organism was Staphylococcus (55%, 11/20), and next was C kroppenstedtii (25%, 5/20). Moreover, we observed only 5.0% of mixed flora grown in the culture of PDM/MDE patients, which is in contrast to the study of Walker et al. The difference of the proportion of predominant bacteria could be explained by the difference in the infective pattern among the different races, living environment, and lifestyles.
The possibility that infectious agent might be involved in IGM has been considered previously. Along with the progress of the microbial detection technology in recent years, the corynebacteria involving granumatous mastitis has been widely concerned. Taylor et al reported that 14 out of the 34 cases had gram-positive bacilli, recognizable as coryneform bacteria, in histological sections. The C kroppenstedtii was also isolated from a breast abscess by the study of Riegel et al. There have been few reports on the mechanisms of IGM caused by Corynebacterium. The first description of NOD2 variant associated with defective neutrophil responses in a woman with granulomatous mastitis regarding Corynebacterium was studied by Bercot and Kannengiesser. In our study, the most common organisms isolated were C kroppenstedtii in 53.7% (22/41). This finding supports that the corynebacteria have an important impact on the bacterial infection of IGM. C kroppenstedtii is difficult to be isolated and identified under the standard conditions of pathology laboratories. The shortest lag time between sampling and bacterial culture, the longest incubation period necessary before the appearance of the first colonies, and the use of a lipid complement are essential factors for bacterial growth and isolation. Isolation is necessary for the successful detection, accurate identification, and antibiotic susceptibility testing of this pathogen. Because Corynebacterium is difficult to isolate and identify, it might be underestimated as the cause of breast abscesses. However, this study is mainly limited by the retrospective study and lacking of specific mechanisms of the mastitis, further study is needed to uncover the pathways for the regulation of inflammation of mastitis.
In conclusion, the median age of PDM/MDE patients is significantly higher than IGM patients. NPM with nipple discharge or inverted nipple symptoms were more likely to be diagnosed as PDM/MDE. The combination of demographic characteristics and clinicopathological characteristics were helpful in differential diagnosis of PDM/MDE and IGM. The elevated serum PRL level may increase the risk for NPM, especially for PDM/MED. The etiology of NPM appears to be localized bacterial infection. Antibiotics to Corynebacterium and Staphylococcus may provide a possible effective therapeutic and preventive strategy.
YDY, SYH, YHL, and BBD participated in research design and the writing of the paper. BBD, QL, WW, YC, and MXZ participated in the performance of the study. YDY and SYH contributed new reagents or analytic tools. QL, BJC, YTJ, SL, and SYH participated in the data analysis. All authors read and approved the final version of the paper.
This work was supported by grants from the National Natural Science Foundation of China (no. 81272897, 81772837), the Science Foundation of Guangdong Province (no. 2016B030229004), and the Sun Yat-Sen Initiative Program for Scientific Research (no. YXQH201701). The funders played no role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Institutional review board statement
The study was approved by the institutional review board of the Breast Tumor Centre of Sun Yat-Sen Memorial Hospital, China and follows the Declaration of Helsinki.
Conflicts of interest
The authors declare that they have no conflicts of interest.
1. Albasri AM. Profile of benign breast diseases in western Saudi Arabia. An 8-year histopathological review of 603 cases. Saudi Med J
2. Michie C, Lockie F, Lynn W. The challenge of mastitis. Arch Dis Child
3. Nemenqani D, Yaqoob N. Fine needle aspiration cytology of inflammatory breast lesions. J Pak Med Assoc
4. Foxman B, D’Arcy H, Gillespie B, et al. Lactation mastitis: occurrence and medical management among 946 breastfeeding women in the United States. Am J Epidemiol
5. World Health Organization. Mastitis. Causes and Management:World Health Organization; 2000.
6. Thomassen P, Johansson VA, Wassberg C, et al. Breast-feeding, pain and infection. Gynecol Obstet Invest
7. Betzold CM. Infections of the mammary ducts in the breastfeeding mother. J Nurse Pract
8. Morrill JF, Pappagianis D, Heinig MJ, et al. Detecting Candida albicans
in human milk. J Clin Microbiol
9. Francis-Morrill J, Heinig MJ, Pappagianis D, et al. Diagnostic value of signs and symptoms of mammary candidosis among lactating women. J Hum Lact
10. Hanavadi S, Pereira G, Mansel RE. How mammillary fistulas should be managed. Breast J
11. Hartley MN, Stewart J, Benson EA. Subareolar dissection for duct ectasia and periareolar sepsis. Br J Surg
12. Dixon JM, Ravisekar O, Chetty U, et al. Periductal mastitis
and duct ectasia: different conditions with different aetiologies. Br J Surg
13. Hughes LE. Non-lactational inflammation and duct ectasia. Br Med Bull
14. Ectasia. PMDPeriductal mastitis duct ectasia. World J Surg
15. Hovanessian Larsen LJ, Peyvandi B, Klipfel N, et al. Granulomatous lobular mastitis: imaging, diagnosis, and treatment. AJR Am J Roentgenol
16. Dursun M, Yilmaz S, Yahyayev A, et al. Multimodality imaging features of idiopathic granulomatous mastitis
: outcome of 12 years of experience. Radiol Med
17. Rahal RM, de Freitas-Junior R, Paulinelli RR. Risk factors for duct ectasia. Breast J
18. Belaabidia B, Essadki O, el Mansouri A, et al. Idiopathic granulomatous mastitis
: apropos of 8 cases and review of the literature. Gynecol Obstet Fertil
19. Diesing D, Axt-Fliedner R, Hornung D, et al. Granulomatous mastitis. Arch Gynecol Obstet
20. Furlong AJ, al-Nakib L, Knox WF, et al. Periductal inflammation and cigarette smoke. J Am Coll Surg
21. Rahal RM, Junior RF, Reis C, et al. Prevalence of bacteria in the nipple discharge of patients with duct ectasia. Int J Clin Pract
22. Ramalingam K, Srivastava A, Vuthaluru S, et al. Duct ectasia and periductal mastitis
in Indian women. Indian J Surg
23. Limaiem F, Korbi S, Tlili T, et al. Idiopathic granulomatous mastitis
mimicking breast cancer: report of two cases. Pathologica
24. Berinder K, Akre O, Granath F, et al. Cancer risk in hyperprolactinemia patients: a population-based cohort study. Eur J Endocrinol
25. Tworoger SS, Eliassen AH, Zhang X, et al. A 20-year prospective study of plasma prolactin as a risk marker of breast cancer development. Cancer Res
26. Foitzik K, Langan EA, Paus R. Prolactin and the skin: a dermatological perspective on an ancient pleiotropic peptide hormone. J Invest Dermatol
27. Lin CH, Ma H, Chung MT, et al. Granulomatous cutaneous lesions associated with risperidone-induced hyperprolactinemia in an amputated upper limb. Int J Dermatol
28. Maier WP, Berger A, Derrick BM. Periareolar abscess in the nonlactating breast. Am J Surg
29. Berná JD. Follicular occlusion due to hyperkeratosis: a new hypothesis on the pathogenesis of mammillary fistula. Med Hypotheses
30. Bundred NJ, Dixon JM, Lumsden AB, et al. Are the lesions of duct ectasia sterile? Br J Surg
31. Walker JC, Sandison AT. Mammary-duct ectasia. A clinical study. Br J Surg
32. Binelli C, Lorimier G, Bertrand G, et al. Granulomatous mastitis and corynebacteria infection. Two case reports. J Gynecol Obstet Biol Reprod (Paris)
33. Taylor GB, Paviour SD, Musaad S, et al. A clinicopathological review of 34 cases of inflammatory breast disease
showing an association between corynebacteria infection and granulomatous mastitis. Pathology
34. Riegel P, Liégeois P, Chenard MP, et al. Isolations of Corynebacterium kroppenstedtii
from a breast abscess. Int J Med Microbiol
35. Bercot B, Kannengiesser CC. First description of NOD2 variant associated with defective neutrophil responses in a woman with granulomatous mastitis related to corynebacteria. J Clin Microbiol
Keywords:Copyright © 2018 The Chinese Medical Association. Published by Wolters Kluwer Health, Inc.
bacteriological etiology; breast disease; idiopathic granulomatous mastitis; mammary ductectasia; periductal mastitis