ABSTRACT
Actinomycosis is a chronic infection caused by Actinomyces species characterized by abscess formation, tissue fibrosis, and draining sinuses. The spectrum of infections caused by Actinomyces species ranges from classical invasive actinomycosis to a less invasive form of superficial skin and soft tissue infection. We present a review detailing all Actinomyces species isolated from breast infections in NHS Lothian between 2005 and 2013, Actinomyces species isolated from breast infections referred to the United Kingdom Anaerobe Reference Unit between 1988 and 2014, and cases describing Actinomyces breast infections published in the medical literature since 1994. Actinomyces species are fastidious organisms which can be difficult to identify and are likely to be underascertained as a cause of breast infections. Due to improved diagnostic methods, they are increasingly associated with chronic, recurrent breast infections and may play a more significant role in these infections than has previously been appreciated.
INTRODUCTION
Actinomycosis is a chronic, invasive, progressive, and often relapsing granulomatous infection caused by Gram-positive, facultatively anaerobic, rod-shaped bacteria belonging to the genus Actinomyces. Classical actinomycosis in humans is typically caused by Actinomyces israelii and is characterized by deep invasive abscess formation, tissue fibrosis, and draining sinuses affecting cervicofacial, thoracic, abdominopelvic areas (1). A number of more recently described Actinomyces species and Actinomyces-like organisms have been associated with less invasive superficial soft tissue infections and are isolated from abscesses at various anatomical sites (2–4).
Breast infections are frequently encountered in primary care and breast clinic settings. They can occur in the parenchyma of the breast or the overlying skin and may be in lactating or nonlactating breasts (5). Lactating breast infections are usually caused by Staphylococcus aureus (6). The microbial etiology of nonlactating breast infections, particularly those which are chronic or recurrent, is more variable, often polymicrobial, and predominantly anaerobic (5, 7). A study of nonpuerperal breast infections has demonstrated that, when culture methods are used which enhance recovery of fastidious anaerobic organisms, almost 25% of recurrent breast abscesses (8 of 33 patients) and 10% of primary breast abscesses (2 of 19 patients) involved Actinomyces species (7). However, primary Actinomyces infection of the breast, first described by Ammentorp in 1893 (8), is generally considered to be rare. A clinical review from 1994 reported 19 cases described in the English language literature (9). Diagnosis of Actinomyces breast infection was often made following surgical intervention, although the method of diagnosis was not specified. Duration of symptoms was reported to range from 1 to 8 years, with two-thirds of the patients presenting with recurrent abscesses. The remaining third of the patients were reported to have examination findings suggestive of malignancy. Most of the patients received extensive surgical treatment, with 11 patients undergoing a mastectomy, presumably for management of infection, as only 3 mastectomies were performed on patients with suspected malignancy. Subsequent to this review, case reports and studies have been published describing a further 27 breast infections caused by Actinomyces species.
Actinomyces breast infection is likely to be underascertained in routine clinical practice, as these fastidious organisms are notoriously difficult and laborious to identify using conventional laboratory methods (2–4). Subsequently, there is a risk that cultures of Actinomyces species are simply identified morphologically as diphtheroids and dismissed as skin commensals, even when grown from an abscess sample. However, new methods of identification, such as matrix-assisted laser desorption ionization–time of flight mass spectrometry (MALDI-TOF MS), allow rapid and reliable identification of many bacteria, including Actinomyces-like organisms (10–12). MALDI-TOF MS and similar technologies are increasingly being adopted by routine diagnostic laboratories worldwide (13).
We present a series describing all cases of Actinomyces species isolated from breast infections at the Edinburgh Breast Unit over an 8-year period from 2005 to 2013 (Table 1). In addition, we include data about Actinomyces species isolated from breast samples which were referred to the Anaerobe Reference Unit, Cardiff, Wales, between 1988 and 2014 (Fig. 1). We then summarize the findings of case reports describing breast infections caused by Actinomyces species published since the review from 1994 (9) (Table 2).
Cases of Actinomyces species isolated from breast infections at the Edinburgh Breast Unit over an 8-year period from 2005 to 2013
Actinomyces species isolated from breast infections referred to the United Kingdom Anaerobic Reference Unit from United Kingdom hospitals, 1988-2014.
Published cases of Actinomyces species isolated from breast infections reported with clinical details since 1994c
(The abstract was presented as a poster at the Association of Breast Surgery Conference in Liverpool, May 2014.)
MATERIALS AND METHODS
Bacterial isolates.Review of electronic records of breast fluid aspirates received between 2005 and 2013 at the microbiology laboratory at the Royal Infirmary of Edinburgh identified 11 cases of breast infections with Actinomyces species and one case with the Actinomyces-like organism Actinobaculum schaali. Specimens were collected as pus in sterile containers or on swabs (Stewart's medium) and routinely transported to the laboratory. Fastidious anaerobic agar with horse blood (not prereduced) was used for culture and incubated in an anaerobic cabinet (80% nitrogen, 10% CO2, 10% hydrogen) for at least 48 h. Until 2011, Gram-positive rods were identified using biochemical methods, generally API Coryne (bioMérieux). From 2011 onward, isolates were identified using MALDI-TOF MS (Bruker Daltonics).
The United Kingdom Anaerobe Reference Unit (UKARU) in Cardiff, Wales, provided details of Actinomyces isolates referred to them from hospitals throughout the United Kingdom between 1988 and 2014 when the source stated on the request form was breast (abscess/fluid/wound).
MALDI-TOF MS.MALDI-TOF MS identification was carried out using a Bruker microflex LT mass spectrometer (Bruker Daltonics) and Bruker flexControl v3.3 software. Isolates were analyzed using a formic acid-based direct, on-plate preparation method. A thin smear of organism was applied to a target plate using a cocktail stick and allowed to dry, and then 1 μl of 100% formic acid was placed on top and allowed to dry. This mixture was overlain with 1 μl of matrix solution (cyano-4-hydroxycinnamic acid) and allowed to dry prior to analysis using the MALDI biotyper. Manufacturer-recommended cutoff scores were used for identification, with p scores of >2.000 indicating identification to the species level, scores between 1.700 and 1.999 indicating identification to the genus level, and scores of <1.700 indicating no identification.
Molecular identification.Definitive molecular identification was by 16S sequencing using the following method. 16S rRNA genes were extracted using a Chelex resin/boiling method and amplified by PCR using the universal primers pA & pH. After purification (Qiaquick PCR purification kit 28106; Qiagen), a second PCR was performed using a primer internal to the initially amplified region (kk) and dye-terminated nucleotides (BigDye 3.1 terminator ready reaction kit). After a second purification step, the sequence of bases was detected by size/dye terminator of the resulting DNA fragments (ABI 3100). The sequences were compared locally with those of other bacteria (ARU bespoke database; Bionumerics, Applied Maths) or with those listed in international databases (NCBI, BLAST), with 16S rRNA gene homology of >97% used to determine bacterial species.
Clinical review.Paper and electronic patient records of cases were reviewed for information on age, smoking history, diabetes, nipple piercing, steroid use, and whether the patient was lactating at the time of infection. The number of times the patient came into contact with the Edinburgh breast unit was recorded along with examination findings and management received. Information on breast infections was collected from general practitioner clinics along with the type, duration, and number of antibiotic courses for breast infection in the community.
Literature review of published cases was completed on Pubmed and Ovid databases using the keywords Actinomyces, actinomycosis, breast, infection, and abscess. Cases of Actinomyces infection of the breast published with clinical descriptions between 1994 and 2013 are detailed in Table 2. Two cases from 1987 not included in the 1994 review (9) are included in this table.
RESULTS
Lothian cases.Over an 8-year period (2005 to 2013), we identified 11 cases of breast infections at our center caused by Actinomyces species (Table 1) and one case with the Actinomyces-like organism Actinobaculum schaali. The predominant Actinomyces species isolated from our subgroup of patients were Actinomyces europaeus (n = 5), A. neuii (n = 3), and A. radingae (n = 3). Identification using MALDI-TOF MS was attempted for 11 isolates, and all of these correlated at the species level with the definitive molecular identification, with p scores ranging from 1.779 to 2.331. Coinfecting organisms were present in half of these cases (n = 6) and usually were unidentified anaerobes. Ten of 12 cases (83%) had chronic, recurrent infections ranging from 2 to 8 episodes (mean, 2.8 episodes), some recurring over many years. Three patients had hidradenitis suppurativa, 6 patients were smokers, 4 patients were diabetic, and 3 patients had a combination of risk factors. No patients in our cohort had a lactational breast abscess, and there was no record of any patients having had a nipple piercing.
Case 5 in particular highlights the difficulties associated with diagnosing and managing Actinomyces breast infections. This patient had 7 episodes of breast infection and abscess formation over a 10-year period, treated with short antibiotic courses. Cultures of aspirated abscess material repeatedly failed to grow organisms or were reported to grow diphtheroids of uncertain significance. Following a positive growth of A. radingae, the patient received a 3-month course of antibiotics and has since had no further relapses (almost 2 years later).
UKARU cases.Over a 26-year period (1988 to 2014), 61 isolates identified as Actinomyces species from breast infections were referred to the United Kingdom Anaerobe Reference Unit (UKARU) from United Kingdom hospitals (Fig. 1). Although Actinomyces species are not considered true anaerobes, the UKARU has developed extensive expertise over many years regarding Actinomyces species. This expertise was driven largely by a referral demand from users for advanced identification of clinically relevant isolates initially categorized as anaerobic Gram-positive rods or anaerobic coryneforms. It is likely that the cases listed here represent only a small proportion of United Kingdom cases, as referral of isolates to the unit is not mandatory. Unfortunately, a further limitation of the referral process is that clinical information is not available for many of these cases; however, a small number (n = 5) state either recurrent or previous breast abscess. One case worthy of particular mention states “recurrent breast abscess for 11 years,” from which A. radingae was isolated.
Literature review cases.Fifteen cases of Actinomyces breast infection were identified on literature review between 1994 and 2013, with another 12 cases (7, 14) found prior to 1994, not included in the review by Jain et al. (9) (Table 2). This paper therefore reports 27 published cases of Actinomyces breast infection in addition to the 19 reported in 1994, although clinical details are only available for 17 of these 27 cases. There was no clearly predominant Actinomyces species. Five cases reported coinfecting anaerobes (17, 24), and one case reported coinfecting Staphylococcus aureus (20), with no mention of coinfecting organisms in 11 of 17 cases. Seven cases were recurrent infections, and a range of different treatment combinations were required to reach abscess resolution. This ranged from 2 to 3 weeks of oral antibiotics to incision and drainage with prolonged antibiotics for 2 to 6 months; the most extreme treatment was tumorectomy of the breast (22). As with our cohort of patients, prior to the diagnosis of Actinomyces breast infection, some patients had been repeatedly treated without success (14, 17, 20).
Combined number and species of Actinomyces breast infections describeda
Combined cases.The Actinomyces species most commonly isolated from breast infections according to a combined table of cases (Table 3) (n = 102) are A. neuii (n = 19), A. europaeus (n = 18), A. turicensis (n = 16), A. radingae (n = 15), and A. odontolyticus (n = 10). These species mostly belong to the group of Actinomyces species generally considered to be less invasive, although it is noteworthy that, in the Lothian and United Kingdom cohorts, A. radingae was isolated from those cases with the greatest number of relapses. The distribution of Actinomyces species broadly reflects previous findings regarding superficial Actinomyces soft tissue infections (2–4), although these studies did not look specifically at breast infections.
Within the NHS Lothian and the published cases, 48% (n = 14) presented clinically with an abscess, 33% (n = 9) presented with a breast mass, 10% (n = 3) presented with a fistula, and 7% (n = 2) presented with periductal mastitis. There was an average of 2.8 episodes of infection per patient in the NHS Lothian cases. Within the published cases, excluding case 1 who had numerous yearly episodes of recurrent infection for 23 years, there was an average of 1.5 episodes of Actinomyces breast infection per patient.
DISCUSSION
Humans and animals are the natural reservoirs of Actinomyces species, which until recently have not been found to exist freely in nature (29). Their normal habitat is the mucosal membranes of the oropharynx, gastrointestinal tract, and female genital tract. They are inherently low in virulence and may rely on the presence of copathogens, such as anaerobic bacteria, to enhance pathogenicity (1). Disruption of the mucosal barrier is the usual precursor to infection with Actinomyces species and, in the breast, the ductal system may serve as a portal of entry. Actinomycosis of the breast usually presents as a chronic, recurrent abscess which in some cases can be difficult to distinguish from inflammatory carcinoma (9, 15, 26). Fistulas and purulent or bloody discharge from sinuses may occur, which may discharge sulfur granules (26). In advanced prolonged cases, fibrosis with architectural distortion of the breast tissue is present on mammography (15).
The pathogenesis and true pathogenic role of various Actinomyces species isolated from breast infections and the treatment required for this has not been clearly defined. This is further complicated by the uncertain etiology of different types of chronic abscess-forming inflammatory conditions involving the breast from which Actinomyces-like organisms can be isolated, such as granulomatous lobular mastitis, hidradenitis suppurativa, and periductal mastitis. Granulomatous lobular mastitis presents as a peripheral inflammatory mass which may simulate malignancy or infection. Patients with this condition often develop multiple and recurrent abscesses. It has been suggested that the Corynebacterium spp. play a part in this condition, (30) but antibiotics effective against these organisms rarely lead to resolution of disease; thus, they may not have a major etiological role. Hidradenitis suppurativa is an inflammatory disease of unclear etiology which commonly affects the axilla and groin and can also affect the skin of the lower half of the breast, resulting in recurrent episodes of abscess formation (31). Recent evidence suggests that anaerobic actinomycetes may be involved in the disease process, especially when lesions are more severe (32). Periductal mastitis is a condition linked to cigarette smoking (33) in which the subareolar ducts are damaged and become infected, often by anaerobic bacteria (34). Women may present with subareolar inflammation, abscesses, and fistulas (35). Smoking has consistently been identified as a risk factor for primary breast abscess and its recurrence (5, 35, 36). Other factors, such as diabetes mellitus, obesity, African-American origin, and nipple piercing, have less consistently been associated with breast abscesses (5, 35).
Despite finding 12 cases over 8 years at our center, which is comparable to the number of cases described in the medical literature over the same time period, we suspect that there were many missed identifications. During the 8-year study period, we found another 15 cases in Lothian where potential Actinomyces-like organisms were isolated from recurrent breast abscesses, but further identification was not attempted and a report was sent out describing diphtheroids of doubtful or uncertain significance. In addition, 4 of our 12 culture-positive cases had previous samples with isolates of potential Actinomyces-like organisms reported as diphtheroids of doubtful significance. This supports the assumption that Actinomyces breast infections may easily go undiagnosed in routine clinical practice. Actinomyces species are slow to grow and notoriously difficult to identify using conventional laboratory methods, often requiring reference laboratory referral for reliable identification. When Actinomyces species do grow on culture, they can resemble other diphtheroid-like Gram-positive rods, such as Corynebacterium species, many of which are considered to be part of normal skin flora. Actinomyces species which are isolated from breast abscess samples may therefore be presumptively identified in the laboratory as diphtheroids based on their morphology and reported as diphtheroids of doubtful or uncertain significance. However, laboratories are increasingly adopting new methods of identification, such as MALDI-TOF MS (13), which allow rapid and increasingly reliable identification of this problematic group of organisms (10–12). Indeed, most of the cases in Lothian were identified after 2012, which is shortly after our laboratory started using MALDI-TOF MS. With 10 cases diagnosed in 2 years of using MALDI-TOF MS compared to 2 cases over 7 years without MALDI-TOF MS, it is clear that ease of identification is a major factor in the increased recognition of Actinomyces breast infections in our clinical setting. The Anaerobe Reference Unit (ARU) has seen a similar increase in the number of isolates referred to them, with more isolates (n = 26) referred to them over the last 3 years of the recorded period than had been referred over the first 20 years (n = 25). Based on information from referring laboratories, this increase is almost certainly driven by an improvement in the identification of Actinomyces species due to increased use of MALDI-TOF MS. Subsequently, laboratories unfamiliar with these organisms refer them to the ARU for confirmation of identification, susceptibility testing, and clinical advice.
Our results support previous findings that Actinomyces species can be reliably identified using MALDI-TOF MS (10–12), with all 12 of our tested isolates identified to the species level, as confirmed by molecular testing. Five isolates were correctly identified to the species level by MALDI-TOF MS despite identification scores only reaching genus-level confidence (ie, a p score of <2.0). This is in keeping with recent evidence suggesting that the cutoff for species-level identification could be reassessed and perhaps lowered to a p score of ≥1.7 for this group of organisms (10, 12).
We have modified the approach to how organisms from breast samples are identified in Lothian. Breast abscess samples now receive anaerobic incubation for 5 days along with prolonged Actinomyces cultures when clinical details mention chronic or recurring infection. Any Gram-positive bacillus growing from a breast abscess sample is identified using MALDI-TOF MS and should no longer be reported as a diphtheroid of uncertain significance without an attempt to identify the organism.
The primary management of breast abscess is drainage along with antibiotic therapy appropriate for the underlying cause of the abscess (31). When Actinomyces species are isolated, longer courses of antibiotics should be considered. Treatment of classical, invasive actinomycosis, typically caused by Actinomyces israelii and to a lesser extent A. gerencseriae, A. meyeri, A. odontolyticus, and A. viscosus/naeslundii (2, 4, 37), involves prolonged antibiotic therapy. Textbooks commonly advise 2 to 6 weeks of intravenous (IV) penicillin followed by 6 to 12 months of oral penicillin or amoxicillin (38). However, there is evidence that shorter antibiotic courses of under 3 months may be sufficient in some cases (39), particularly those caused by less-invasive Actinomyces species, such as A. europaeus, A. funkei, A. neuii, A. radingae, and A. turicensis (2–4, 37, 40). Even shorter 7- to 14-day courses of oral antibiotics are typically used when treating breast infections, but this is likely to be insufficient for Actinomyces-associated breast infections, and longer courses, in addition to surgical drainage, may be required to prevent recurrences. Actinomyces species are susceptible to many beta-lactam antibiotics, with penicillin and amoxicillin generally regarded as first-choice options (38, 41). However, due to the common presence of coinfecting, beta-lactamase-producing organisms, treatment options should ideally include beta-lactamase-stable antibiotics, such as amoxicillin plus clavulanic acid (41), at least for the initial 2 weeks of treatment. Alternative agents for patients with penicillin allergy could include doxycycline or clindamycin, although there is less evidence for their efficacy (38). We suggest at least 6 weeks of antimicrobial treatment for extensive infections involving Actinomyces species or in cases of recurrence. Although some recurrent cases of Actinomyces breast infections seem to have benefited from this treatment approach, it is not clear whether a longer course of antimicrobial in the first instance would have prevented relapses in these cases.
In conclusion, Actinomyces-associated breast infections are problematic, difficult to diagnose, and difficult to treat. They are increasingly recognized in clinical practice, most likely due to a combination of increased awareness and improved diagnostic methods. Further studies are required to clarify the pathogenic role of Actinomyces species in various inflammatory conditions which involve the breast, such as periductal mastitis, hidradenitis suppurativa, and granulomatous mastitis. These conditions all present with clinical features similar to those seen in Actinomyces breast infections. Taking into account the fastidious nature of Actinomyces species, it is quite possible that anaerobic actinomycetes are present significantly more frequently than they are found in these conditions. In particular, it is important to clarify whether longer initial courses of effective antibiotic treatment may prevent recurrences and radical surgery when Actinomyces species are isolated in association with these conditions.
FOOTNOTES
- Received 22 April 2015.
- Returned for modification 15 May 2015.
- Accepted 16 July 2015.
- Accepted manuscript posted online 29 July 2015.
- Copyright © 2015, American Society for Microbiology. All Rights Reserved.
