Background

Pneumonia is an important cause of medical morbidity and mortality worldwide; pneumonia and influenza are listed as the 8th leading cause of death in the United States, with 57,062 deaths in 2015 [1]. Well-recognized and common causes of community-acquired pneumonia (CAP) in adults include bacteria such as Streptococcus pneumoniae, Haemophilus influenzae, and Staphylococcus aureus, and viruses such as influenza virus, respiratory syncytial virus, human metapneumovirus, and parainfluenza virus. Intensive recent investigations have failed to identify a causative organism in 50–62% of patients hospitalized for CAP [2, 3].

The most common laboratory report for sputum samples submitted for microbiological study is “mixed respiratory flora.” Pneumonia caused by S. pneumoniae is thought to result from unrecognized microaspiration of pneumococci that have colonized the nasopharynx [4]. The authors of the present study have previously hypothesized that unrecognized aspiration of less virulent bacteria that normally inhabit the nasopharynx might also cause pneumonia, especially in persons whose upper airways are bypassed or whose ability to clear aspirated organisms is damaged [5]. The present study examined the possibilities that: (i) careful examination of Gram-stained sputum and culture plates may reveal a predominant bacterium such as Corynebacterium spp. that are not generally regarded as a pulmonary pathogen; (ii) quantitative cultures may document a high concentration of these bacteria in sputum; and (iii) these organisms are in fact the cause of some cases of pneumonia.

Corynebacterium spp. Gram-positive bacilli that are often dismissed as contaminants have been reported to cause pneumonia (Table 1). New technologies, such as matrix-assisted laser desorption ionization time of flight (MALDI-TOF) mass spectrometry enable rapid and precise identification of bacterial species [6], and highly sensitive techniques such as polymerase chain reaction (PCR) can help exclude a role for viruses, mycoplasma, and chlamydia. With the advent of these new technologies, there has been greater interest in Corynebacteria, and these organisms have been cited as emerging pathogens [7].

Table 1 Previously documented cases of Corynebacterium pulmonary infection

The present study reports on 3 well-documented cases of CAP due to Corynebacteria (2 cases of C. propinquum, 1 case of C. striatum) that were diagnosed in a 6-month period at a large tertiary care hospital. The study reviews the medical literature on Corynebacterium spp. as etiologic agents of pneumonia, with the goal of increasing awareness of the possibility that so-called “normal respiratory flora” may actually be responsible for an as-yet undetermined proportion of cases of CAP.

Case presentations

Case 1

A 75-year-old man on hemodialysis for end-stage renal disease was hospitalized for cough, fever, and altered mental status. His temperature was 103 °F, pulse 68, blood pressure 124/69, respiratory rate 22, and O2 saturation 96% on 2 L of O2 by nasal cannula. He was lethargic, with bibasilar crackles and a normal cardiac examination. His white blood cell (WBC) count was 20,100 cells/mm3 and his plasma procalcitonin was 1.1 ng/mL. Chest X-ray showed bibasilar infiltrates. Microscopic examination of Gram-stained sputum showed profuse polymorphonuclear leukocytes (PMNs) and Gram-positive rods, many of which appeared to be intracellular (Fig. 1). Sputum culture yielded overwhelmingly predominant Corynebacteria, with rare P. aeruginosa. The Corynebacteria was identified as C. propinquum by MALDI-TOF. Blood cultures and viral PCR were negative. The patient was initially treated with vancomycin, cefepime, and metronidazole. Based on the predominance of Corynebacteria with the absence of other bacteria on Gram stain and a negative PCR for respiratory viruses, his pneumonia was attributed to C. propinquum, and only vancomycin was continued. He responded and was discharged to complete a 10-day course of linezolid.

Fig. 1
figure 1

Gram-stain observed in sputum from patient with Corynebacterial pneumonia. A sputum sample collected from the patient in Case 1 depicts Gram-positive bacilli (purple) within and surrounding polymorphonuclear cells (pink). The “club-shaped” appearance pathognomonic for the group “Coryneform” is appreciated

Case 2

A 61-year-old man was found unresponsive post laryngectomy and tracheostomy for laryngeal squamous cell carcinoma. His temperature was 97.8 °F, pulse 99, blood pressure 114/86, and respiratory rate 22. Breath sounds were decreased bilaterally. WBC count was 7200 cells/mm3, lactate 1.2 mg/dL. Chest X-ray showed a new left-sided infiltrate. Sputum Gram stain revealed many Gram-positive rods, including those found within PMNs. Sputum yielded many Corynebacteria, identified by MALDI-TOF as C. propinquum, and few S. aureus. Sputum was liquefied with 2% N-acetyl cysteine and diluted serially; aliquots were cultured on blood agar and the number of colony forming units (CFU) was calculated [8]. The specimen contained 2 × 109 CFU of Corynebacteria and < 105 CFU of S. aureus per mL. Blood cultures and viral PCR were negative. The patient was initially given vancomycin, cefepime, and ampicillin; treatment was switched to ampicillin/sulbactam after the Corynebacterium was reported susceptible. His mental status rapidly improved, and he was subsequently discharged to complete a 10-day course of amoxicillin/clavulanate.

Case 3

A 59-year-old male with widely metastatic squamous cell carcinoma of the tongue was admitted for bleeding from his tracheostomy site. His temperature was 99 °F, blood pressure 108/84, pulse 119, respiratory rate 24, and oxygen saturation 82% on room air. He had blood at the tracheostomy site and bibasilar rhonchi. WBC count was 22,000 cells/mm3, hemoglobin 11.1 g/dL, and lactate 1.8 mg/dL. Chest X-ray revealed a left upper-lobe infiltrate. Sputum sample showed profuse PMNs and Gram-positive rods. Culture yielded C. striatum (confirmed by MALDI-TOF) and few Escherichia coli. The patient was placed on comfort care and died 8 days later.

Discussion

Epidemiology

Non-diphtheria Corynebacterium spp.--commensal flora of the skin, respiratory tract, and mucous membranes [9, 10]-- have been characterized as emerging pathogens [7]. Among reported cases, infections of the respiratory tract are second in frequency only to those of the urinary tract [11]. The literature search revealed 67 reported cases of pulmonary infection due to non-diphtheria Corynebacterium spp., summarized in Table 1. Cases include pneumonia (community-acquired, hospital-acquired, or ventilator-associated), necrotizing pneumonia, and empyema, although pneumonia predominated. Implicated Corynebacterium spp. include C. pseudodiphtheriticum [12, 13], C. propinquum [14], C. striatum [14,15,16,17,18], C. afermentans [19], C. xerosis [20], C. jeikeium [10, 21,22,23], C. pseudotuberculosis [24, 25], C. mucifaciens [26], C. ulcerans [27], and C. macginleyi [28].

Microbiology, pathogenesis, and immunity

“Coryneform bacteria” is the umbrella term for the class of bacteria comprised of facultative intracellular, non-spore-forming, catalase-positive, non-acid-fast, non-motile, and irregularly-shaped Gram-positive rods [29]. Based on 16S rRNA sequencing studies, Corynebacterium belongs to a subdivision of Gram-positive eubacteria with high guanine-to-cytosine content, which also contains Arthrobacter, Mycobacterium, Nocardia, and Streptomyces [30]. Coryneform bacteria are pleomorphic throughout their life cycle, and may have thickenings at either end, giving the “coryneform” name, meaning “club-shaped”; they may be clustered, forming shapes that resemble a “V,” “palisades,” or “Chinese letters” [29]. Because of their relation to C. diphtheriae, their common name is “diphtheroids.”

Little is known about the virulence factors of the non-diphtheria Corynebacterium spp. that may contribute to CAP and other infections [31]. C. pseudodiphtheriticum, the species most commonly implicated as a cause of pneumonia, has an ability to invade human epithelial type 2 (HEp-2) cells and survive for 24 h after infection, suggesting a potential mechanism to escape killing by the innate immune system [31]. Invasive strains form biofilms and bind fibrinogen and fibronectin, which may be facilitated by fimbrial subunits [31, 32].

Predisposing conditions

With the exception of 2 cases of pneumonia due to C. pseudotuberculosis in veterinary workers [24, 25], all other reported cases of Corynebacterial pneumonia have occurred in persons whose upper airways are bypassed, whose ability to clear aspirated organisms is damaged, or who have immunocompromising conditions (Table 2). The great majority of patients have had conditions that affect pulmonary clearance, including chronic obstructive pulmonary disease (COPD), cystic fibrosis [33], and previous radiotherapy to the thorax [27]. Some have had procedures that bypass protective mechanisms of the upper airways such as recent endotracheal intubation [28, 34, 35] or laryngectomy [36], as seen in Cases 2 and 3 in the present study. Among patients who are immunocompromised, co-morbidities included HIV infection [17, 37, 38], chemotherapy [10, 28], hematologic malignancies or conditions that compromise PMN function such as diabetes mellitus, alcohol use disorder [16] or end-stage renal disease, as seen in Case 1 in the present study.

Table 2 Risk factors in reported cases of Corynebacterial pneumonia

Clinical description

The presentation of pneumonia due to Corynebacteria does not differ from that due to bacterial pneumonia of any other cause: cough, fever, and altered mental status are usually present, as observed in the present study’s 3 cases. Cases of community-acquired or ventilator-associated pneumonia, cavitary pneumonia, empyema, and bacteremia have been described (Table 1). Only 5 (7.5%) of the 67 reported patients with Corynebacterial pneumonia have been bacteremic with Corynebacteria. This suggests that blood cultures may be of minimal utility for confirming any sputum findings. Pneumonia caused by the quintessential pulmonary pathogen, S. pneumoniae, is bacteremic in no more than approximately 25% of cases [39], and the rate of bacteremia is far lower in pneumonia due to less pathogenic organisms, such as nontypable H. influenzae [40] or Moraxella catarrhalis [41].

Laboratory diagnosis

Given the limited utility of blood cultures in diagnosing Corynebacterial pneumonia, diagnosis requires a multifaceted approach. In most previous case reports, diagnosis has been based on microscopic examination of a Gram-stained sputum and careful examination of culture plates [42]. In the 3 cases discussed in this report, the sputum Gram stain showed profuse numbers of Corynebacteria. In 12 of the previously reported cases, quantitative studies were carried out on sputum, bronchoalveolar lavage samples, or protected bronchial brush samples (Table 1) with few or no colonies of other potentially pathogenic bacteria, as in Case 2 in which quantitative sputum culture revealed 1.2 × 109 CFU/mL. Pneumonia due to Corynebacteria may occur far more commonly than has been reported, because the finding of these organisms is usually dismissed as “normal mouth flora” or “normal respiratory flora.”

MALDI-TOF is useful in providing precise speciation, as was employed in the 3 cases discussed in this report. An analysis of 83 Corynebacterial samples compared MALDI-TOF with API Coryne V2.0 (bioMérieux, Marcy l’Etoile, France), which is the method that has been most widely used to speciate Corynebacteria in the past few decades. Identification matched in 73 of 83 (88%) samples [43]. 16S rRNA sequencing was used to resolve the discrepancy in the remaining 10 samples. All 10 matched with MALDI-TOF’s identification, which demonstrates the superiority of this methodology [43, 44].

Management

Corynebacterium spp. that have been identified as causes of pneumonia or empyema are delineated in Table 1. C. pseudodiptheriticum, the cause of about two-thirds of reported Corynebacterial pneumonia cases, is susceptible to penicillin, cephalosporins, and vancomycin, but resistant to macrolides, clindamycin, trimethoprim/sulfamethoxazole, quinolones, and/or rifampin (Table 3). Some species, including C. jeikeium, C. macginleyi, and C. xerosis, are resistant to all common antibiotics except glycopeptides such as vancomycin (Table 3). Intermediately resistant organisms such as C. propinquum and C. striatum, implicated in the present study’s 3 cases, are also consistently susceptible to vancomycin and ampicillin, but variably susceptible to penicillin and cephalosporins, and commonly resistant to macrolides, clindamycin, trimethoprim-sulfamethoxazole, fluoroquinolines, and rifampin. The other reported Corynebacterium spp.—C. accolens, C. afermentans, C. mucifaciens, C. pseudotuberculosis, C. urealyticum, and C. ulcerans—had only 1–2 case reports each, and/or did not include comprehensive susceptibility reports. These cases are summarized in Table 3. Although only one pneumonia case report [17] included susceptibility to linezolid or tigecycline, a prospective study analyzing infections due to Corynebacteria from a variety of sites revealed that all isolates—the majority being C. amycolatum, C. accolens, C. minutissimum, and C. glucuronolyticum (strains that are not regularly implicated in pneumonia)—were susceptible to linezolid and tigecycline [45].

Table 3 Antibiotic susceptibilities of Corynebacterium spp. identified as causes of pulmonary infection

These results suggest that, when a Gram-stained sputum from a patient with pneumonia reveals Corynebacteria as the predominant isolate, vancomycin should be the initial antibiotic of choice. If MALDI-TOF identifies a species that is likely to be susceptible to other antibiotics, the empiric antibiotic of choice can be revised, or definitive antibiotic treatment can be given based on results of antibiotic susceptibility testing as determined by gradient diffusion strip testing. There is no accepted recommendation for length of treatment for Corynebacterial pneumonia. However, among case reports, treatment time has ranged from 10 days for a case of C. macginleyi pneumonia [28], to 8 weeks for a cavitating pneumonia due to C. jeikeium [22], to 14 months for pneumonia due to C. pseudotuberculosis [25]. It seems reasonable to select a duration of treatment based on the nature of the lung lesion. Infiltrates should respond to the same 7 days of treatment that are used for other bacterial pneumonias, whereas cavitating pneumonia should be treated for longer periods, perhaps until the cavity is closed. In some cases, inadequate responses have been observed in cases of a lung abscess or cavitation, prompting surgical resection [16, 37].

Conclusions

Although traditionally regarded as part of normal respiratory flora, Corynebacteria has the capacity to cause pneumonia in immunocompromised patients, or in immunocompetent patients who have impaired airway clearance and/or structural damage to the airways. The presentation is similar to that of any other bacterial pneumonia. Such infections may be far more common than has been previously suggested, because the finding of Corynebacteria on Gram stain and culture is likely to lead to a report of “mixed” or “normal respiratory flora.” A correct diagnosis can be made primarily by visualization of large numbers of diphtheroids in microscopic fields that contain > 20 WBC per epithelial cell, especially if the bacteria are located within PMNs, and supported by a culture in which the overwhelmingly predominant organism is Corynebacteria. Speciation by MALDI-TOF is increasingly available in clinical laboratories. Initial treatment should be with vancomycin. Speciation may guide antibiotic choice, but definitive therapy should be based on susceptibility testing using gradient diffusion strip testing methodology.