Corynebacterium kroppenstedtii Is an Emerging Cause of Mastitis Especially in Patients With Psychiatric Illness on Antipsychotic Medication.

This retrospective study of patients with Corynebacterium kroppenstedtii infections revealed a predominance of mastitis and a potential association with psychiatric illnesses. At least one third of our patients with C kroppenstedtii mastitis had psychiatric illness, and >92% received antipsychotic medications. Drug-induced hyperprolactinemia may be an important modifiable risk factor in these patients.

Corynebacterium kroppenstedtii is a lipophilic Corynebacterium first described in 1998 after isolation from a sputum specimen [1]. Association with breast conditions, especially recurrent granulomatous mastitis (GM), was noted subsequently [2-4]. Occasional reports of bacteremia and a single report of prosthetic valve infection have also been described [2, 5]. However, knowledge on risk factors associated with C kroppenstedtii, clinical presentations, and significance remained incomplete. In this study, we analyzed the clinical, microbiological, and investigative findings from 42 patients with C kroppenstedtii.

METHOD

The database of a reference laboratory in Hong Kong was searched retrospectively for C kroppenstedtii isolated from clinical specimens between January 2009 and January 2015. The isolates were identified as C kroppenstedtii by phenotypic characteristics [1], matrix-assisted laser desorption/ionization-time-of-flight mass spectrometry ([MALDI-TOF MS] Bruker Daltonics, Bremen, Germany) MALDI Biotyper version 3.1 with reference library version 5.0.0.0 plus an in-house-enhanced database against C kroppenstedtii [6], and an in-house one-tube nested quantitative polymerase chain reaction (qPCR) targeting C kroppenstedtii-specific sialidase gene, nanI, using primers (CKOut-F-TCGACGTTGAAGACGCCACCACCATCACCGAT, CKOut-R-GGGGGCATCGGCGATGGTGAGTTCCT; CKIn-F-CCAATGGCACAGCATCG, CKIn-R-GAACGTCGGCAAGAAATGC) and CKPb (FAM-CACCAAGTTTTATCTCGACGGGTACCA-3IABkFQ) (GenBank accession no. ACR18588) [7]. The target, nanI, is specific for C kroppenstedtii with no in silico or in vitro cross-reactivity against other Corynebacterium spp. Additional information on laboratory identification is available in the Supplementary Material. Clinical data were retrieved from case records and hospital information system for analysis.

Statistical Analysis

Mann-Whitney U test was used to compare patients’ data as appropriate. A P value of <.05 is considered statistically significant. All statistical analysis was performed using SPSS version 20.

RESULTS

From 2009 to 2015, 46 strains of C kroppenstedtii were isolated (42 patients), including 41 breast-related samples and 5 nonbreast-related specimens. The age ranged from 20 to 90 (median 39 years), and 39 were female. Good identification to species (score of ≥2.0) and genus level (score ≥1.7 to <2.0) was noted in 28 and 7 isolates by MALDI-TOF MS, respectively, and all 46 isolates were nanI qPCR positive.

Corynebacterium kroppenstedtii in Breast-Related Specimens

The 41 breast-related C kroppenstedtii strains (37 patients) included 27 aspirates/biopsies and 14 wound swabs. Corynebacterium kroppenstedtii was found as pure culture in 39 specimens; 1 breast aspirate collected from a lactating woman and 1 swab of nipple discharge from a 5-month postpartum patient yielded additional scanty Acinetobacter sp and coagulase-negative Staphylococcus, respectively. Three patients have multiple specimens, collected at least 1 month apart, yielding C kroppenstedtii.

Histological analyses were available from 29 of the 37 patients. Noncaseating granulomatous inflammation was observed in 11 patients, whereas nonspecific inflammatory changes were noted in others. Granulomatous inflammation was found in patients with longer duration of symptoms compared with patients with nongranulomatous inflammation, although the difference was not statistically significant (median, 30 vs 14 days; P = .099). Age, underlying conditions, clinical presentation, treatment, and recurrences were not significantly different between the 2 groups.

Nine of the 11 patients with GM were investigated for acid-fast bacilli (AFB) (smear for AFB in 4 patients; smear and culture for AFB in 5 patients), and all were negative for AFB. Smear and culture for AFB were also performed in 13 of the 26 patients with non-GM, which were again negative.

Clinical characteristics of the 37 patients are summarized in Table 1. The only patient presented without breast lesions was a breastfeeding mother with left breast mastitis, where C kroppenstedtii was isolated from pus expressed from her nipple.

Table 1.

Clinical Characteristics of Patients With Corynebacterium kroppenstedtii Isolated From Breast-Related Specimens

Characteristics	N = 37 (All Female)
Age (median, range), years	36, 20–52
Clinical Presentation
Laterality
Left	16/37
Right	15/37
Bilateral	6/37
Duration of symptoms before presentation (median), days	2–180 (30)a
Palpable breast abnormality	35/36 (97%)
Pain	16/36 (44%)
Discharge	5/36 (14%)
Laboratory Test Results		Reference Range
White cell count (range, median), ×109/L	6.6–18 (9.6)	3.89–9.93
Absolute neutrophil count (median), ×109/L	2.5–13.9 (7.2)	2.01–7.42
Lymphocyte count (median), ×109/L	1.2–3.4(1.9)	1.06–3.61
Past History
History of recurrent abscesses	10/35 (29%)
Diabetes/impaired glucose tolerance	5/35 (14%)
Hypertension	2/35 (6%)
Pituitary tumor (macro- or microadenoma)	4/35 (11%)3 with evidence of hyperprolactinaemia
Psychiatric history	13/35 (37%)
Schizophrenia	9
Bipolar affective disorders	3
Substance abuse with drug-induced psychosis	1
Psychiatric medication	12/12b
Pregnancy history
Pregnant	1 (3rd trimester)
Within 12 months postpartum	3
Management
Antibiotics	31/34
Aspiration only	13/33
At least 1 I&D	13/33
Aspiration and I&D	4/33
Resection of lesion	3/33
Steroids	0/34
Outcome
Recovered	14/20 (duration of follow-ups: 1 to 12 months, median 3 months)
Recurrence	6/20

Abbreviations: I&D, incision and drainage.

aData available from 32 patients.

bNumber of patients on medications with known association with hyperprolactinemia: 6 risperidone, 3 haloperidol, 2 chlorpromazine, 2 fluoperazine, 1 flupenthixol, 1 olanzapine, 1 paroxetin. Number of patients on medications with no known association with hyperprolactinemia: 8 trihexophenidyl, 1 aripiprazol, 1 clonazepam. Eleven patients are on more than 1 psychiatric medications.

NOTE: Denominators denote the number of patients with details available for analysis.

Past medical and psychiatric histories were available from 35 of the 37 patients. Four patients had pituitary tumor. Thirteen patients (37%) were diagnosed with a psychiatric illness, with schizophrenia being the commonest. Drug history was available from 12 of the 13 patients, and all regimens included antipsychotic drugs associated with hyperprolactinemia. Two patients with schizophrenia had concurrent pituitary tumor.

History of antibiotic use and other interventions were available from 34 and 33 patients, respectively. Exact dates of antibiotics were available in 15 patients, where 5 patients were receiving antibiotics at the time, or within 1 week before specimen collection. Overall, β-lactam antibiotics were the most commonly used antibiotics. At least 15 patients underwent more than 1 invasive intervention. Follow-up data were available from 20 of the 37 patients, and 3 of the 6 patients with recurrence had a psychiatric illness.

Corynebacterium kroppenstedtii in Other Specimens

Corynebacterium kroppenstedtii were isolated from 3 blood cultures, 1 Tenchkoff exit site swab, and 1 sputum. Patients with C kroppenstedtii bacteremia were 80, 82, and 90 years old, 2 were male, and all presented with low-grade fever with no obvious foci. They received ceftriaxone, vancomycin, and ampicillin-clavulanate, respectively. All 3 patients recovered fully. Corynebacterium kroppenstedtii was also isolated from Tenckhoff exit site in a 31-year-old male patient with Tenckhoff exit site infection. He recovered after 3 weeks of oral ampicillin. One 64-year-old male had dry cough, weight loss, and consolidative lung lesion despite multiple courses of antibiotics including levofloxacin, piperacillin-tazobactam, and metronidazole. The clinical significance of C kroppenstedtii in his sputum specimen was doubtful.

DISCUSSION

To the best of our knowledge, our study included the largest cohort of patients with C kroppenstedtii infection. The predominance of breast-related specimens (89%) is consistent with previous reports of its potential association with mastitis and abscesses [2–4, 8].

Several additional findings were noted. First, psychiatric illnesses were seen in 37% of patients with C kroppenstedtii-related mastitis/abscesses. This is significantly higher than the local prevalence of (1) 8.97% in the female population with mixed anxiety and depressive disorders and (2) 2.5% of psychotic disorders in Hong Kong [9, 10]. These patients also appeared to have a higher risk of recurrence of mastitis/abscesses. Although C kroppenstedtii-related GM has been described in 2 patients with psychiatric illnesses [11], our finding involving more patients put forward a cogent argument for the potential association between the 2. We suspected that antipsychotic-induced hyperprolactinemia is the unsuspected culprit, which was also postulated in the case report [11]. Hyperprolactinemia has also been reported in patients with “idiopathic” GM [12, 13]. It is likely that some of these idiopathic GM were C kroppenstedtii-related but were misdiagnosed as idiopathic because routine culture methods do not reliably isolate this slow-growing organism. Prolactin was thought to induce ductal ectasia and milk stagnation, predisposing patients to GM. Moreover, the proinflammatory effect of prolactin was observed in bovine mammary epithelial cells, through induction of nuclear factor κ-light-chain enhancer of activated B-cell signaling pathway [14]. It is unfortunate that serum prolactin level was not performed in most of our patients. Although further research is required to establish the exact role of hyperprolactinemia in C kroppenstedtii-related mastitis/abscesses, we believe that prolactin level should be routinely checked in these patients, to identify patients who may benefit from withdrawal of prolactin-inducing drugs. Besides drug-induced hyperprolactinemia, other potential contributing factors, eg, behavioral and personal hygiene, cannot be excluded and deserve further investigations.

Second, a spectrum of histopathological appearance is observed in our patients. Granulomatous inflammation was seen more frequently in patients with prolonged symptoms, implying a possible common etiology between nongranulomatous and GM.

Our study is limited by the retrospective nature of data retrieval. First, investigations to exclude other GM-associated pathogens, in particular dimorphic fungi and Mycobacterium tuberculosis (MTB), were not performed in all patients. Nonetheless, histoplasmosis and blastomycosis previously related to GM are extremely rare in our locality. For MTB, of the 11 patients with GM potentially suggestive of MTB, smear or culture of AFB was performed in 9 patients, and all yielded negative results. The only patient with subsequent right cervical tuberculous lymphadenitis was believed to be irrelevant, because 2 years lapsed between the resolution of left breast abscess and MTB lymphadenitis. Second, exact dates of antibiotics were only available in 15 patients. Hence, we cannot assess whether antibiotic use shortly before or at specimen collection could have selected towards isolation of C kroppenstedtii. Nonetheless, a metagenomic analysis of microbiota in breast specimens taken before antibiotic treatment also revealed a predominance of Corynebacterium spp, especially C kroppenstedtii, among patients with nonlactating GM [15]. Their finding supported that C kroppenstedtii plays a role in GM. Thirdly, prolactin level was not tested in most of our patients; hence, the role of drug-induced hyperprolactinemia in C kroppenstedtii-related mastitis and psychiatric illness remained speculative. Finally, potential diagnostic bias of GM in women with psychiatric illness cannot be excluded in our study. Future study comparing prevalence of psychiatric illnesses in patients with C kroppenstedtii and non-C kroppenstedtii mastitis/abscesses is required.

CONCLUSIONS

In conclusion, our finding strengthened the association of C kroppenstedtii with breast diseases. Further study is warranted to confirm the association among C kroppenstedtii, mastitis/abscesses, and psychiatric illnesses. Finally, we propose that serum prolactin level should be routinely checked in all patients with C kroppenstedtii-related mastitis/abscesses.

Supplementary Data

Supplementary materials are available at Open Forum Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.

Click here for additional data file.