Nonlactational Mastitis in Western Brazilian Amazon Mastitis Cases

ABSTRACT.

Chronic mastitis varies in etiology between its lactational and nonlactational forms and can be challenging to diagnose. This study aimed to assess the epidemiological profile and diverse etiologies of nonlactational mastitis in Amazonas, Brazil, focusing on distinguishing between tuberculous mastitis (TM) and idiopathic granulomatous mastitis (IGM). This is a retrospective and prospective study that was carried out at the mastitis outpatient clinic of Fundação de Medicina Tropical Doutor Heitor Vieira Dourado from 2013 to 2021 and evaluated epidemiological data, imaging, and laboratory tests. Descriptive statistics were performed. In this retrospective and prospective analysis, 124 medical records were initially considered, with 12 excluded for various reasons. The remaining 112 cases underwent thorough evaluation through epidemiological data, imaging, and laboratory tests, by employing descriptive statistics for analysis. The pathology revealed a predominant prevalence of IGM (64.3%), followed by various forms of mastitis, including confirmed TM (4.5%), presumable TM (8.9%), and others. Our findings indicate that IGM, though a rare cause of breast masses and abscesses, accounts for a significant portion of mastitis cases. Histopathological studies were essential for diagnosis, with ultrasound being the primary imaging tool. This study is one of the largest Brazilian series on nonlactational mastitis, highlighting the condition’s complexity and diverse manifestations in the Amazon region.

INTRODUCTION

Chronic inflammatory mastitis, which can be lactational or nonlactational, presents significant challenges for diagnosing the underlying etiology and making therapeutic evaluations.¹ Among the types of nonlactational mastitis, idiopathic granulomatous mastitis (IGM) or granulomatous lobular mastitis is particularly notable. Pathologically, IGM is characterized by sterile, noncaseating, lobulocentric granulomatous inflammation.²

In cases of lactational mastitis, infection is the common underlying cause, and Staphylococcus aureus is commonly implicated as the sole pathogen. However, in nonlactational mastitis, a polymicrobial origin is observed in approximately 30% of cases. Tuberculous mastitis (TM) and nontuberculous mycobacteriosis mastitis often caused by Corynebacterium spp. are rare etiologic agents but can be underlying causes of IGM.^3–6

It is especially important to consider TM in differential diagnoses in areas of endemicity such as the Brazilian Amazon.⁷ It is rare in the breast and represents less than 0.1–2.5%^1,8,9 of cases of extrapulmonary manifestation, which is influenced by the prevalence of tuberculosis,¹⁰ with rates of 3–4.5% in developing countries.¹¹

The diagnostic challenge is exacerbated by limitations within the Brazilian public health system, particularly in terms of diagnostic evaluation and care flow. This study aims to evaluate the epidemiological profile of nonlactational mastitis in the state of Amazonas. It seeks to differentiate between IGM and TM and to elucidate the varied etiologies of this condition through findings obtained from ultrasound imaging.

Materials and Methods

This study, both retrospective and prospective in nature, focused on patients with nonlactational mastitis at the outpatient clinics of Fundação de Medicina Tropical Dr. Heitor Vieira Dourado (FMT-HVD) in Manaus, Amazonas, Brazil, from 2013 to 2021. The Research Ethics Committee of Universidade Federal do Amazonas approved the study (CPEA 20093719.8.0000.5020).

Patients selected for this study had a minimum of two outpatient follow-ups, during which either a secretion culture was collected or a histopathological breast examination was conducted. Medical records were accessed via the iDoctor^® system, and data were compiled using REDCap^® software (version 9.1.0). Data collection involved a standardized questionnaire covering epidemiological, etiological, and imaging variables, along with clinical progression. Key data points included patient demographics, history of current illness, clinical symptoms, risk factors, and clinical findings. Special attention was given to diagnostic delay, defined as more than 2 months from symptom onset to service presentation, and factors such as previous empirical antibiotic therapy and diagnostic methods (fine needle aspiration [FNA], core needle biopsy, purified protein derivative [PPD], Ziehl-Neelsen [ZN] staining, culture, polymerase chain reaction, erythrocyte sedimentation rate, and histopathological study).

The routine service process involved FNAs being sent for laboratory analysis, including acid-alcohol-resistant bacillus, rapid molecular test, and culture testing for mycobacteria, along with mycology and aerobic bacterial cultures. Secretions, when present, were also sent for mycobacterial culture. Samples were transported to the laboratory in Styrofoam boxes with ice.

Imaging tests evaluated included mammography, breast ultrasound, magnetic resonance imaging (MRI), chest X-rays, and computed tomography, as available in the medical records. Data on first consultation, diagnosis, last consultation, and diagnostic evolution were also recorded.

Criteria for presumable TM (P-TM) were established because of the difficulty in confirming TM. These criteria encompassed clinical presentation, PPD results, histology indicative of IGM or nonspecific chronic mastitis, response to therapy, and absence of recurrence posttreatment for a minimum of 6 months.

Statistical analysis was conducted using χ² tests, Fisher’s test for variables with counts under 5, and the Mann-Whitney test for continuous variables. IBM SPSS^® for Mac^® v. 20 was used for these analyses.

RESULTS

In this study, we initially analyzed 124 medical records from female patients. Twelve of these records were excluded: six because of the absence of histopathologic studies and six because of duplication in the retrospective and prospective data sets.

The patients’ ages ranged from 21 to 68 years, with an average of 36.7 years. When examining risk factors, we found that 1.4% of patients with IGM had a history of tuberculosis, compared with 17.2% in the TM/P-TM group, a difference that was statistically significant. None of the patients in either group tested positive for HIV.

The clinical presentation of the cohort predominantly included symptoms such as pain, erythema, nipple discharge, and fever and/or chills. Regarding breast complaints, 49 (68.1%) of the patients presented with a nodule or mass, and 46 (63.9%) presented with an abscess. Other findings included fistulae, nipple discharge, axillary adenopathies, and areolar retraction. Areolar retraction showed a statistically significant difference in occurrence. Table 1 presents the clinical findings.

Table 1.

History and clinical evaluation of nonlactational mastitis patients

Variables	IGM, n/N (%)	TM/P-TM, n/N (%)	P-Value
Risk factors
HIV	0/72 (0.0)	0/15 (0.0)	–
Previous tuberculosis	1/72 (1.4)	2/15 (17.2)	0.043
Side
Right	31/72 (43.1)	7/15 (46.7)	0.798
Left	28/72 (38.9)	5/15 (33.3)	0.687
Bilateral	13/72 (18.1)	3/15 (20.0)	0.860
Place
Multiple	47/72 (65.3)	5/15 (33.3)	0.090
Upper quadrants	26/72 (36.1)	5/15 (33.3)	0.838
Lower quadrants	2/72 (2.8)	3/15 (20.0)	0.009
Central	6/72 (8.3)	2/15 (13.4)	0.177
Medical history
Pain	68/72 (94.4)	14/15 (93.3)	0.827
Erythema	62/72 (86.1)	15/15 (100.0)	0.308
Nipple discharge	39/72 (54.2)	9/15 (60.0)	0.852
Fever and chills	32/72 (44.4)	8/15 (53.3)	0.530
Axillary adenopathy	7/72 (9.7)	4/15 (26.7)	0.072
Breast history
Nodule/mass	39/72 (54.2)	10/15 (66.7)	0.375
Abscess	37/72 (51.4)	9/15 (60.0)	0.543
Diffuse	7/72 (9.7)	2/15 (13.3)	0.676
Clinical examination
Fistulas	31/72 (43.1)	8/15 (53.3)	0.467
Nipple discharge	24/72 (33.3)	8/15 (53.3)	0.144
Axillary adenopathy	18/72 (25.0)	4/15 (26.7)	0.893
Areolar retraction	3/72 (4.2)	3/15 (20.0)	0.028

IGM = idiopathic granulomatous mastitis; P-TM = presumable tuberculous mastitis; TM = tuberculous mastitis.

There was no statistically significant difference regarding side when IGM and TM/P-TM were compared; the main location of mastitis in patients with information in the medical records was multiple locations, followed by locations in the upper quadrants and lower quadrants in two patients with IGM and three with TM/P-TM. There was a significant difference in lower quadrant involvement.

In terms of laboratory and imaging evaluations, smear microscopy showed alterations in only 7.1% of patients, whereas culture results were positive in 1.8%. Ultrasound, conducted for all patients, revealed abscesses in 43 (59.7%), hypoechoic nodules in 32 (44.4%), fistulae in 27 (37.5%), skin thickening and edema in 11 (15.3%), confluent nodules in 10 (13.9%), and axillary adenopathy in 10 (13.9%) (Table 2). Mammography was performed on 23 patients, with alterations observed in 4 cases. Magnetic resonance imaging was conducted for one patient and showed no significant abnormalities.

Table 2.

Imaging exams of nonlactational mastitis patients

Category	IGM, n/N (%)	TM/P-TM, n/N (%)	P-Value
Ultrasound
Abscess	36/72 (50.0)	7/15 (46.7)	0.814
Hypoechoic nodule	29/72 (37.7)	3/15 (20.0)	0.138
Fistula	21/72 (29.2)	6/15 (40.0)	0.409
Skin thickening and swelling	6/72 (8.3)	5/15 (33.3)	0.008
Confluent nodules	9/72 (12.5)	1/15 (6.7)	0.519
Axillary adenopathy	9/72 (12.5)	1/15 (6.7)	0.519
Mammography with alterations	3/72 (4.2)	1/15 (6.7)	0.640

IGM = idiopathic granulomatous mastitis; P-TM = presumable tuberculous mastitis; TM = tuberculous mastitis.

Pathological examinations indicated that 64.3% of patients had IGM, 4.5% had confirmed TM, and 8.9% had P-TM (Table 3). Other diagnoses included nonspecific chronic mastitis (6.3%), nontuberculous mycobacteriosis mastitis (2.7%), GM of fungal origin (2.7%), and rare cases caused by Escherichia coli, Staphylococcus epidermidis, and Corynebacterium spp. Polymerase chain reaction testing was performed on 50 patients considered to have nonlactational mastitis, with 12 testing positive and 38 testing negative. Only one patient exhibited radiologic alterations suggestive of pulmonary tuberculosis.

Table 3.

Frequency of histopathology/etiology findings of infectious mastitis

Histopathologic diagnostics	n/N (%)
GM
IGM	72/112 (64.3)
TM	5/112 (4.5)
P-TM	10/112 (8.9)
P-TM with unusual criteria	9/112 (8.0)
Nonspecific chronic mastitis	7/112 (6.3)
Non-TM	3/112 (2.7)
Specific
Fungal	3/112 (2.7)
Escherichia coli	1/112 (0.9)
Staphylococcus epidermidis	1/112 (0.9)
Corynebacterium spp.	1/112 (0.9)

GM = granulomatous mastitis; IGM = idiopathic granulomatous mastitis; Non-TM = nontuberculous mastitis; P-TM = presumable tuberculous mastitis; TM = tuberculous mastitits.

DISCUSSION

In general, lactational mastitis is treated by obstetricians, and nonlactational and recurrent mastitis are referred for more specialized evaluations. However, both can have infectious and noninfectious etiologic agents. In the first presentation, they are characterized by the formation of an abscess, and treatment with empirical antibiotic therapy is usually performed, the main etiological agents being S. aureus, Streptococcus agalactiae, and Mycoplasma spp. In the absence of resolution via antibiotic therapy, patients are submitted to a breast biopsy, followed by a histopathological evaluation, which may have influenced the high rate of diagnostic delay in this series. A biopsy is important given the need to exclude the presence of an associated neoplasm,^12–14 assists in the etiological assessment, and is performed in cases of noninitial resolution. In the literature, although GM or tuberculous GM^15,16 can mimic a neoplasm, their association with neoplasms is rare.¹⁷ In this study, we observed that both lactational and nonlactational mastitis can be caused by infectious and noninfectious etiologies. This aligns with existing literature, emphasizing the complexity of diagnosing and treating mastitis. Notably, our findings show a high prevalence of breast abscesses and nodules, corroborating the literature which suggests that nonlactational mastitis often presents with these symptoms.

Imaging exams help to exclude other pathologies; however, they are nonspecific regarding the etiology and depend on the anatomopathological exams.¹⁸ We found that ultrasound, widely regarded as the diagnostic tool of choice for mastitis, was the most frequently used imaging modality in our study (Figures 1 and 2). This is consistent with previous research indicating its utility in identifying heterogeneous areas and atypical lymph nodes, which are common in mastitis but can also mimic breast cancer.^19,20 The Doppler ultrasound can demonstrate increased vascularization in and around the lesions.²¹ Mammography is usually performed in cases with longer evolution and atypical conditions and in younger women. Breast MRI is rarely performed,²⁰ though it helps in the differential diagnosis of neoplasia, and there may be hyperintensity of the T2 signal due to edema and peripheral enhancement of abscesses with kinetic curves usually of types 1 and 2.^2,21–23 Although MRI is of limited value, it can be used to assess the response to therapy.^24,25 Interestingly, in our study, MRI was seldom used, reflecting its limited application in mastitis outside the context of neoplasia differentiation.

Figure 1.

Patient diagnosed with fungal mastitis (Candida krusei). (A) Image of the left breast demonstrating a suspected area of mastitis prior to treatment; (B) image demonstrating left breast without significant skin changes after 4 years; (C) ultrasound image of the breast demonstrating focal ectasia of the left breast duct; and (D) ultrasound image demonstrating a cyst in the left breast.

Figure 2.

Patient diagnosed with breast tuberculosis. (A) Ultrasound image demonstrating retroareolar duct dilatation with thick material inside the right breast; (B) ultrasound image demonstrating accumulation of thick fluid in the right breast; (C) image demonstrating skin retraction in the patient’s breast in the third month of tuberculosis treatment; and (D) image demonstrating skin retraction in the patient’s breast on the fifth month of tuberculosis treatment.

In our study, a high prevalence of breast abscesses was observed, followed by nodules with circumscribed margins in the images. Our clinical and radiological findings corroborate the literature,²⁰ with ultrasound being the test of choice, and no case was observed in which MRI was performed.

This study represents the largest Brazilian series to evaluate the clinical and epidemiological characteristics of nonlactational mastitis.⁴ In the literature, nonlactational GM is usually described in retrospective case series with a long evaluation time; however, despite this situation, the number of patients is often limited.^18,20,26,27 Another aspect to consider is that in the presence of GM, the etiological agent should be sought, and only if the agent is unknown should IGM be considered, which represents most cases. A study carried out at two referral centers in Hong Kong and China evaluated 102 patients with IGM, with a mean age that was similar to that of the patients in our study and most of whom had painful masses during clinical examination, which was a finding that was also prevalent in our study. In addition, biopsies were performed in 63 patients, and the majority (76.2%) had no bacterial growth. Of the 15 species of bacteria found, growth of species such as Corynebacterium, S. aureus, and Streptococcus²⁸ was observed. An Indian study evaluating GM cases over a 5-year period initially observed 68 patients with mastitis, 12 patients being excluded because of lactational origin and 6 patients because of lack of information. Of the remaining 50 patients, 38 had IGM and 12 had periductal mastitis.¹ Comparing our findings with other studies, we noticed similarities in the clinical presentation of mastitis. For instance, studies in Hong Kong, China, and India also reported high occurrences of painful masses and abscesses, mirroring our observations. These similarities reinforce the global nature of mastitis characteristics despite geographical differences.

Most cases of GM are represented by IGM, and a smaller portion has proven etiology.⁴ In a systematic review of nonlactational infectious mastitis in the Americas, it was observed that 84.8% of the analyzed patients had bacteria as the etiological agent, with tuberculosis in 38.4% in addition to fungal agents in 10.5%, viral agents in 1.2%, and parasitic agents in 1.2%.⁴ Care should be taken when evaluating the numbers, since in our series, 64.3% were considered to have IGM, and only 4.5% had confirmed TM, 8.9% had P-TM, and 8.0% had P-TM without criteria, a fact also considered in the literature, due to the difficulty in obtaining a definitive diagnosis.²⁹ If we consider only proven cases, our series matches the literature^11,30; however, if we consider probable cases, GM would represent 13.4% of cases, a frequency that is rare and reported only in developing countries.^31,32 In the Brazilian Amazon, the incidence of tuberculosis is 74.1 cases per 100,000 inhabitants,⁷ and a low incidence was observed in this study, which is probably due to the difficulty involved in etiological diagnosis, either via histopathological study including ZN staining or via culture.³³

From a clinical point of view, GM cases clinically present as breast masses and abscesses. The diagnosis can be made only via a histopathological study, which is characterized by multinucleated giant cell granulomas with microabscesses. Inflammatory breast carcinoma and other infectious and noninfectious causes, such as tuberculosis, other bacterial infections and those of fungal or parasitic origin, sarcoidosis, Wegener’s granulomatosis, giant cell arteritis, polyarteritis nodosa, and foreign body reaction, should be excluded.^22,34 Granulomatous mastitis cases are usually unilateral and may involve all quadrants or be associated with mammary duct ectasia and Corynebacterium infection.²⁷ In our study, we had one case of GM caused by this bacterium.

Our study contributes significantly to the understanding of nonlactational mastitis in the Brazilian Amazon, a region with specific epidemiological characteristics. The lower incidence of tuberculosis-related mastitis in our study compared with the general prevalence in the Amazon highlights the potential regional variations in etiological factors.

However, our study faced limitations. The retrospective design and lack of a standardized imaging evaluation protocol might have introduced variability in our data. Moreover, the difficulty in achieving a specific diagnosis, particularly in cases of IGM and TM, reflects a broader challenge in mastitis research.

In conclusion, our study underscores the diversity of nonlactational mastitis in the Brazilian Amazon. It highlights the need for continuous research for more accurate and timely diagnoses to prevent significant morbidities. Developing specific referral protocols and diagnostic procedures, particularly in specialized centers, could greatly aid in the management of mastitis.