Idiopathic Granulomatous Mastitis in Hispanic Women—Indiana, 2006-2008 FREE

MMWR. 2009;58:1317-1321

1 figure, 2 tables omitted

Idiopathic granulomatous mastitis (IGM) is a rare inflammatory breast lesion of unknown etiology that occurs in women of childbearing age; only a few hundred cases have been reported worldwide.^1- 2 A breast cancer mimic, IGM also is diagnosed by breast biopsy.³ On December 12, 2008, a physician in Indianapolis, Indiana, reported a cluster of seven IGM diagnoses in multigravid Hispanic women, an unusually high number. To evaluate the etiology of the breast masses and characterize associated epidemiologic and clinical features, the Indiana State Department of Health and CDC conducted a multidisciplinary investigation. This report describes the results of that investigation. A total of nine cases of IGM were confirmed during 2006-2008 in Indianapolis, for an annual prevalence of IGM of 2.4 per 100,000 women aged 20-40 years. The prevalence was 12 times higher among Hispanic women. Among IGM patients at the hospital, a median of 5 months elapsed between symptom onset and diagnostic biopsy. Histopathologic evaluations confirmed IGM. In a case-control study of all seven cases and 21 controls from the hospital, case-patients were significantly more likely than controls to have less than a sixth-grade education (odds ratio [OR] = 12.7), a positive tuberculin skin test (OR = undefined), or a medication allergy (OR = 15.0). No other risk factors were significantly associated with case status. Barriers to accessing health care, including low education level, resulted in delayed care for breast masses. Future research could provide more complete descriptions of the epidemiology and etiology of IGM.

The hospital is a university-associated, county medical center that emphasizes care of vulnerable populations.* Before this cluster, the reporting physician had never seen a case of IGM in 15 years of practice. All seven women had sought medical care for painful breast masses and received a diagnosis of IGM based on pathologic findings during biopsies to rule out malignancy. The masses were palpable (ranging from 1 × 1 cm to 6 × 4 cm in size) and unilateral in the left breast in six cases and bilateral in one case. All seven patients reported at least two pregnancies and had breastfed at least one of their children. Symptoms occurred a median of 34 months (range: 10-62 months) after last pregnancy and 26 months (range: 4-46 months) after cessation of breastfeeding. The median time between symptom onset and biopsy was 5 months (range: 3-6 months); the patients attributed the delays to barriers in accessing health care, including concerns about financial implications, occupational repercussions, and/or child care obligations. After symptom onset, case-patients missed or cancelled 23% of their appointments. One patient was deported before completing treatment.

All seven biopsy specimens were sent to CDC for additional histopathologic evaluation.† This testing confirmed IGM in all seven patients, each with noncaseating granulomas, acute and chronic inflammation, and absence of foreign body material in breast biopsy tissue. Eosinophilia was noted in specimens from five patients. Special stains revealed no evidence of mycobacteria, spirochetes, other bacteria, fungi, or trypanosomes. Immunohistochemistry (IHC) on the tissue specimens for mycobacteria and Trypanosomacruzi were negative. IHC using polyclonal antibodies for Corynebacteriumdiphtheriae showed rare epithelial staining in four of seven cases, although this test has unknown specificity. However, polymerase chain reaction (PCR) assays on DNA extracts from the formalin-fixed tissues using broad-range panbacterial 16S rDNA primers were negative.

The seven case-patients reported no major commonalities in workplace or type of work; foods or products sent from abroad; vitamins, supplements, diet products, or local herbal medicines⁴; personal hygiene and grooming products; fertility treatment history or contraception used; schools, churches, grocery stores, or laundromats attended; hobbies and recreational activities; breast trauma; or sexual practices involving the breasts. No patient reported animal exposure or bites, ingesting unpasteurized cheese or milk, tattooing or nipple piercing, or injections or cosmetic treatments of any kind. None currently used alcohol, tobacco, or other drugs.

At the time of interview in January 2008, two case-patients reported that their symptoms had resolved, and five reported ongoing relapses. After receiving a diagnosis of IGM, five patients had received short-duration antibiotics (7-14 days of a penicillin, cephalosporin, clindamycin, sulfonamide, and/or metronidazole). Three had received surgical incision and drainage procedures. None had received steroids. Of the seven, six had received a tuberculin skin test (TST) within 3 years of their IGM diagnosis, and five were positive (10-18 mm); at least four previously had received Bacille Calmette-Guerin (BCG) vaccination. Follow-up testing was nondiagnostic for active tuberculosis,‡ and no patient received antitubercular medications. Routine single-drug tuberculosis prophylaxis would have been recommended for the five case-patients with a positive TST result, but none attended their scheduled follow-up visits.

A case-control study was conducted to identify possible risk factors for disease. A case was defined by a diagnosis at the hospital of an idiopathic granulomatous reaction of the breast by tissue pathology during 2000-2008. Three unmatched controls per case were selected randomly from among all Hispanic women aged 20-40 years treated at the hospital's outpatient clinics during 2008. All case-patients agreed to be interviewed; 39 women were contacted to obtain 21 controls. CDC and Indiana State Department of Health clinicians abstracted data from inpatient and outpatient medical records to clarify demographics, comorbidities, and clinical courses. In-depth ethnographic interviews with cases and controls were conducted by a two-person CDC team of female, bilingual Spanish speakers using a standard questionnaire. Statistical testing was conducted using the Fisher's exact test, two-sided, with statistical significance defined as p<0.05. All interviews were conducted in Spanish.

Cases and controls were comparable on most demographic characteristics, including age, birthplace, years residing in the United States, and number of pregnancies and births. Case-patients were significantly more likely than controls to report less than a sixth-grade education (OR = 12.7). Case-patients were also significantly more likely than controls to have a known allergy to medication such as sulfa-containing drugs (OR = 15.0) or a positive TST result (OR = undefined), although significantly fewer controls reported any TST result (OR = 12.0). Self-defined breastfeeding difficulties, including engorgement and pain, were common in both groups (OR = 2.1) and not significantly different.

The Hispanic population is the fastest growing demographic group in Indianapolis, increasing by an estimated 70% during 2000-2007.§ To estimate the overall rate of IGM in the city's population, investigators conducted citywide case finding using standard search terms (“breast” and “granuloma,” or “breast” and “granulomatous,” followed by individual record review to confirm idiopathic etiology) to query the specimen and diagnostic fields of all three pathology system databases servicing the major medical hospitals in Indianapolis. A total of nine confirmed case-patients from three hospitals, including all seven from the initial cluster, were diagnosed between January 1, 2006, and December 31, 2008; eight were Hispanic women and one was a non-Hispanic white woman. Prevalence was calculated using demographic estimates from the U.S. Census Bureau. The annual prevalence of IGM during 2006-2008 in Indianapolis was 2.4 per 100,000 women aged 20-40 years and 28.3 per 100,000 Hispanic women of the same age range.

M Goldman, MD, HM Selke, DO, I Pardo, MD, SE Clare, MD, PhD, RE Emerson, MD, Indiana Univ School of Medicine; JF Howell, DVM, Indiana State Dept of Health. W-J Shieh, MD, PhD, S Zaki, MD, PhD, Div of Viral and Rickettsial Diseases, National Center for Zoonotic, Vector-Borne, and Enteric Diseases; C Sanchez, RL Sinkowitz-Cochran, MPH, A Srinivasan, MD, M Jhung, MD, Div of Healthcare Quality Promotion, National Center for Preparedness, Detection, and Control of Infectious Diseases; TJ Chester, MD, Career Epidemiology Field Officer Program, Coordinating Center for Terrorism Preparedness and Emergency Response; M Ritchey, DPT, VG Jarquin, PhD, E Meites, MD, EIS officers, CDC.

This is the largest cluster of idiopathic granulomatous mastitis reported in the United States. This is also the first cluster reported among a particular ethnic group residing in the United States. Previous reports from other countries have suggested that IGM is more prevalent among nonwhite women worldwide.^3,5- 6 Based on this investigation, the prevalence of IGM in Indianapolis was 12 times higher among Hispanic women than among all women aged 20-40 years in the city. Like all IGM, the causes for the cases in this cluster are still unclear. The patients described in this report all had in common low education levels, difficulty communicating in English, and substantial barriers to accessing health care, resulting in long delays in receiving medical care for their breast symptoms. Granulomatous scarring of the breast might have resulted as an uncommon late manifestation of one or more separate disease processes (e.g., undiagnosed or untreated allergic reactions or infections) that might have been identifiable if treated earlier.

IGM is an exceedingly rare diagnosis with a variable clinical course and no clear etiology. It is a diagnosis of exclusion, made after malignancy and known granulomatous etiologies (e.g., mycobacterial infections, sarcoidosis, and fungal and other infections) have been ruled out.⁵ Although no evidence-based therapeutic guidelines exist, attempted treatment modalities have included antibiotics, steroids, surgery, and more recently, disease-modifying antirheumatic agents (e.g., methotrexate).^7- 8 Although IGM is nonmalignant, illness can last years, even with treatment, and symptoms can be debilitating.^5,7- 8

Previously published reports have proposed possible etiologic contributors to IGM including breastfeeding practices, breast trauma, smoking, oral contraceptive use, hypersensitivity reactions to breast secretions, or Mycobacterium, Corynebacterium, or other infections.^6- 10 The findings of the investigation described in this report did not clearly support any of these hypotheses. Case-patients were significantly more likely than controls to have a history of allergic reaction to medication, and eosinophilia was present in five of seven biopsies; these findings hint at an immune-mediated reaction, although no clear precipitator of an allergic or hypersensitivity reaction was identified. Because all seven patients were born and raised in countries where tuberculosis is endemic and BCG vaccination is routine, the five positive TST results, with no other signs or symptoms of active tubercular disease, probably represented past exposure to Mycobacterium species or residual reaction to BCG, rather than active mycobacterial infection isolated to the breast; however this unlikely etiology could not be completely excluded because of the difficulty of laboratory confirmation of fastidious species. Next, the epithelial immunostaining pattern observed with polyclonal antibodies to C.diphtheriae suggests that bacteria cross-reacting to these antibodies contribute to pathogenesis (even though polyclonal antibody IHC results often are nonspecific); however, the negative PCR results with panbacterial 16S rDNA did not support this hypothesis. Finally, histopathologic examination found no convincing evidence of other acute infections.

The findings in this report are subject to at least three limitations. First, selecting all Hispanic controls limited the ability to assess the role of ethnicity. Second, higher participation rates among cases than controls (100% versus 54%) might have introduced bias, potentially masking actual differences. Finally, the power of this small study to identify risk factors associated with IGM was low.

Delayed care for breast masses is especially concerning among women with IGM, in part because of the need to rule out other disease promptly (e.g., breast cancer). When an idiopathic condition such as IGM is identified, a thorough clinical history and evaluation should be conducted to exclude known causes of granulomas. Future research could provide more complete information about the etiology and epidemiology of IGM, including possible contributions of ethnicity and socioeconomic status.

This report is based, in part, on contributions by M Rosenman, MD, Regenstrief Institute, Inc.; KL Clayton, OW Cummings, MD, CFM Danielson, MD, PhD, AM Freeman, PJ Garrison, JD, L Harris, MD, C Meece, S Spinola, MD, S Wilson, MD, Indiana Univ School of Medicine; JA Monroe, MD, L Robertson, MS, Indiana State Dept of Health; and T MacCannell, PhD, and J Perz, DrPH, Div of Healthcare Quality Promotion, National Center for Preparedness, Detection, and Control of Infectious Diseases, CDC.

Idiopathic granulomatous mastitis (IGM) is an exceedingly rare inflammatory breast lesion of unknown etiology.

What is added by this report?

This is the largest cluster of IGM reported in the United States; all affected patients were young Hispanic women who experienced delays in receiving health care.

What are the implications for public health practice?

Future research could provide more complete information about the epidemiology and etiology of IGM, including possible contributions of ethnicity and socioeconomic status.

*In 2009, 39% of patients treated at the hospital were uninsured, 32% were covered by Medicaid, and 18% were covered by Medicare. Since 1995, the hospital has operated a Hispanic Health Services office, including language interpretation and scheduling assistance, to serve a growing number of Hispanic immigrants. In 2008, a total of 31,450 women aged 20-40 years had at least one visit to the hospital system; 7,095 (23%) were Hispanic (Regenstrief Institute, Inc., unpublished data, 2009).

†Histopathologic evaluation included Gram stain, acid-fast bacillus (AFB) smear, Grocott's methenamine silver stain, Warthin-Starry stain, and immunohistochemistry testing.

‡Five women received chest radiographs; all were clear except for one with a single calcified granuloma. Five had AFB smears, and four had AFB cultures performed on their biopsy specimens, all of which were negative.

§U.S. Census Bureau 2005-2007 American Community Survey 3-year estimates. Available at http://factfinder.census.gov.