Severe Pneumonia Caused by Corynebacterium striatum in Adults, Seoul, South Korea, 2014–2019

Most (70.4%) cases were hospital-acquired, and 51.9% of patients were immunocompromised.

C orynebacterium striatum is a nonlipophilic, fermentative coryneform bacterium that commonly occupies the normal flora of the skin and oropharynx (1). Although C. striatum isolated from clinical specimens has frequently been considered a contaminant, it is increasingly recognized as a pathogen of various infections, including central line-associated bacteremia (2), endocarditis (3), and pleuropulmonary infection (4)(5)(6). In 1980, C. striatum was reported as a cause of pleuropulmonary infection in a patient with chronic lymphocytic leukemia (4). In 2018, a group of researchers in the United States reported 3 cases of community-acquired pneumonia (CAP) in which Corynebacterium species were the predominant isolate and suggested that Corynebacterium species are a noteworthy clinical cause of pneumonia (6). However, scarce information is available on the incidence, clinical characteristics, and outcomes of severe C. striatum pneumonia in critically ill adult patients, because previous studies included <5 patients with severe C. striatum pneumonia, except those reporting hospital outbreak events.
Methicillin-resistant Staphylococcus aureus (MRSA) is a major cause of severe hospital-acquired pneumonia (HAP), and the clinical characteristics and outcomes of severe MRSA pneumonia are well-documented. Therefore, comparing C. striatum and MRSA pneumonia could clarify the clinical characteristics of C. striatum pneumonia for clinicians. We investigated the proportion, clinical characteristics, and outcomes of severe C. striatum pneumonia in adults and compared those aspects with those for severe MRSA pneumonia.

Study Design, Setting, Data Collection, and Patient Selection
This study is part of an ongoing prospective observational study on severe pneumonia in critically ill adult (>16 years of age) patients at Asan Medical Center, a 2,700-bed tertiary referral center in Seoul, South Korea. Since March 2010, we have prospectively identified all adult patients admitted to the 28-bed medical intensive care unit (ICU) who were clinically suspected of having severe pneumonia and monitored them until hospital discharge (7)(8)(9)(10). We collected data on patient demographics; underlying diseases or conditions; category of pneumonia; initial clinical manifestations; laboratory, microbiologic, and radiologic findings; treatment; complications; and mortality rates. For this study, we investigated patients with severe C. striatum pneumonia who were admitted to the medical ICU during January 2014-December 2019. This study was approved by the institutional review board of Asan Medical Center (IRB no. 2010-0079), which waived the need for informed consent due to the observational nature of the study.

Definitions
We defined and categorized pneumonia as previously stated (11)(12)(13). We defined severe pneumonia as the necessity for mechanical ventilation or having septic shock at ICU admission (12). We defined sepsis and septic shock according to Sepsis-3 criteria (14). We defined immunocompromised state as described previously (15).

C. striatum Identification and Antimicrobial Susceptibility Testing
We cultured sputum specimens on a 5% sheep blood plate and MacConkey agar (Synergy Innovation, http:// www.synergyinno.com). When coryneform gram-positive bacilli were isolated, we identified and performed antimicrobial susceptibility testing for specimens that were urea positive or from the ICU (16). We quantitatively cultured bronchoalveolar lavage specimens on chocolate agar and identified and performed susceptibility testing when coryneform gram-positive bacilli exclusively grew at >10 4 CFU/mL (16). Until August 2015, our facility used the triple sugar iron, motility, API Coryne (bioMérieux-Vitek, https://www.biomerieux. com) system to identify coryneform gram-positive rods. In September 2015, our facility began using matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (Bruker Daltonik, https://www.bruker. com). We determined antimicrobial susceptibility profiles by ETEST (bioMérieux-Vitek) with MHF medium (Mueller-Hinton agar with 5% horse blood + 20 mg/L β-NAD; bioMérieux-Vitek). We used the Clinical and Laboratory Standards Institute M45 guideline for interpreting susceptibility test results (17) and defined multidrug resistance as resistance to >3 antimicrobial drug families.

Statistical Analysis
We compared patient demographics, underlying diseases and conditions, and clinical and laboratory parameters between the C. striatum group and the MRSA group. We used χ 2 or Fisher exact test to compare categorical variables and Student t-test or Mann-Whitney U test to compare continuous variables. We analyzed changes in the proportions of pneumonic pathogens over time by using a χ 2 test for trend. We performed all analyses in SPSS Statistics 24.0 (IBM Corp., https://www.ibm.com) and considered p<0.05 statistically significant.

Demographics, Underlying Diseases and Conditions, and Pneumonia Categories
During the study period, we identified a total of 1,740 patients with severe pneumonia. Among them, 27 had severe C. striatum pneumonia and 103 had severe MRSA pneumonia ( Table 1). The median patient age in the C. striatum group was 72.0 years and in the MRSA group was 71.0 years. Solid cancer, diabetes mellitus, and structural lung diseases were the most common underlying conditions in both groups. More patients in the C. striatum group were immunocompromised (51.9% vs. 26.2%; p = 0.01). Most (70.4%) patients in the C. striatum group had HAP, 14.8% had healthcare-associated pneumonia (HCAP), 11.1% had ventilator-associated pneumonia, and 3.7% had CAP. HAP was significantly more common in the C. striatum group than the MRSA group (70.4% vs. 42.7%; p = 0.01); HCAP was more common in the MRSA group (32.0% vs. 14.8%; p = 0.08), albeit without statistical significance.

Clinical Manifestations and Laboratory Findings
Dyspnea, fever, sputum, and cough were the most common signs and symptoms in both groups (Table  4). Fever tended to be less common in the C. striatum group (66.7% vs. 82.5%; p = 0.07). The proportion of patients with septic shock at the time of ICU admission was significantly higher in the MRSA group (67.0% vs. 44.4%; p = 0.03). However, the proportion of mechanical ventilation, acute physiology and chronic health evaluation (APACHE II) score, and sequential organ failure assessment (SOFA) score at the time of ICU admission were similar between the 2 groups. Peripheral leukocyte counts, platelet counts, and serum C-reactive protein levels also were similar between the 2 groups, but serum procalcitonin level was significantly higher in the MRSA group than the C. striatum group (median 0.3 ng/mL vs. 1.8 ng/mL; p<0.01). grade moderate (3+), 1 grade few (2+), and 3 were grade rare (1+) (Appendix Table, https://wwwnc. cdc.gov/EID/article/28/11/22-0273-App1.pdf). All 27 C. striatum isolates underwent antimicrobial susceptibility testing. All isolates were resistant to penicillin, ceftriaxone, erythromycin, and ciprofloxacin, and susceptible to vancomycin, and all isolates were multidrug resistant.

Outcomes
The mortality rates between the C. striatum and MRSA group showed no statistically significant differences: 30-day mortality (40.7% vs. 29.1%; p = 0.25), 60-day (48.1% vs. 42.7%; p = 0.61), and 90-day (59.3% vs. 50.5%; p = 0.42) ( Table 5). In-hospital mortality rates were higher (70.4%) in the C. striatum group than in the MRSA group (52.4%), albeit without statistical significance (p = 0.09). Mortality rates were similar for C. striatum and MRSA in subgroups regardless of the patient's immune status. We noted no statistically significant differences in the median length of ICU stay between the C. striatum and MRSA group, both 14 days (p = 0.33), nor in the length of hospital stay after ICU admission, 30 days for the C. striatum versus 29 days for the MRSA group (p = 0.48).

Discussion
We investigated the proportion and characteristics of severe C. striatum pneumonia compared with severe MRSA pneumonia. Although the proportion of severe MRSA HAP greatly decreased during 2014-2019, the proportion of severe C. striatum pneumonia sharply increased and surpassed that of severe MRSA pneumonia. C. striatum pneumonia was more commonly associated with immunocompromise, viral co-infection, and fungal co-infection. Mortality rates between the C. striatum and MRSA groups were comparable. We found that the proportion of severe MRSA pneumonia decreased while severe C. striatum pneumonia greatly increased and that C. striatum emerged as one of the most common pathogens in patients with severe HAP. Strengthened infection control measures during the study period might have contributed to the decline of severe MRSA pneumonia (18); however, severe C. striatum pneumonia demonstrated the opposite trend. Several possible explanations for this discrepancy exist. First, detection of C. striatum from respiratory specimens in clinical laboratories increased, possibly because experience among laboratory staff accumulated over time. Also, new reliable identification techniques, such as matrix-assisted laser desorption/ionization time-of-flight mass spectrometry, were introduced and enabled precise and rapid detection and identification of bacteria in clinical samples, which might have contributed to the increased reports of severe C. striatum pneumonia (19,20). Second, C. striatum can be resistant to infection control measures and can adhere to abiotic surfaces and form biofilms on various medical devices, such as feeding tubes, endotracheal tubes, and ventilators (21,22). Some reports documented C. striatum strains with resistance to high-level disinfectants, such as 2% glutaraldehyde and other biocides (23,24). These findings suggest that appropriate environmental infection control measures for C. striatum should be further investigated and implemented. Finally, hospital outbreaks also might have contributed to the seeming discrepancy. Colonized patients and contaminated inanimate objects could be reservoirs for prolonged outbreaks. However, when we chronologically analyzed the occurrence patterns according to time and place, we could not find any suggestions of notable outbreaks. Clinical observation alone creates difficulties and limitations in distinguishing outbreaks; therefore, future studies should include more detailed   typing analysis of C. striatum isolates to identify and curb possible healthcare-associated outbreaks. In this study, viral or fungal co-infection was more common in the C. striatum group, whereas other bacterial co-infection was more common in the MRSA group. This finding could represent the host factor because a greater proportion of C. striatum patients were in an immunocompromised state, which would make them vulnerable to opportunistic infections. Of note, fewer cases of bacterial coinfection were diagnosed in the C. striatum group, but the cause for this difference is uncertain. One possible explanation is that C. striatum might influence the behavior and fitness of other bacteria. A recent study reported that Corynebacterium species can reduce the toxicity of Staphylococcus aureus by exhibiting decreased hemolysin activity and displaying diminished fitness of in vivo coinfection (25). Further targeted studies on this issue are needed. We found that serum procalcitonin level was higher in the MRSA group than in the C. striatum group (median 1.8 ng/mL vs. 0.3 ng/mL). Some studies suggest that serum procalcitonin can be used as a marker for bacterial infection and to differentiate bacterial from viral infection or noninfectious causes of inflammation (26,27). In 2017, a group of researchers in China reported that the median serum procalcitonin level of an S. aureus bacteremia group of patients was higher (1.18 ng/mL) than that of a coagulase-negative staphylococci bacteremia group (0.21-0.31 ng/mL) (28). We speculate that infections caused by low-virulence bacteria, such as C. striatum in our study, might have low levels of procalcitonin and this warrants further investigation.
Mortality rates were similarly high in both groups, but septic shock at the time of initial clinical manifestation was less common in the C. striatum group. Immunocompromised conditions were more common in the C. striatum group, which could suggest that C. striatum is less virulent than MRSA. Host factor might contribute to the development of severe C. striatum-associated pneumonia and the subsequent outcomes; however, we noted no statistically significant differences in mortality rates between the 2 groups after stratification by immunocompromised conditions. The existence of co-infection and pathogen types (e.g., other bacteria, viruses, fungi) involved might have affected mortality rates, but we were unable to effectively evaluate each effect because of the small number of patients in each subgroup.
The first limitation of our study is that we used a single-center design and our results might not be replicable in other centers or hospital systems. In addition, as we mentioned previously, we were not able to effectively evaluate the sole contribution of C. striatum because co-infection with other pathogens was common among the patient cohort. Finally, we included all C. striatum isolates from sputum, endotracheal aspirate, and bronchoalveolar lavage, but the cultures were mostly semiquantitative, and some of the C. striatum isolates might have been nonpathogenic colonizers. A 2020 study from the United States reported that normal respiratory flora appears to have caused one quarter of CAP cases (29), which supports our finding that bacteria previously considered as colonizers or normal flora can be a cause of pneumonia.
In conclusion, we found C. striatum was associated with severe HAP. Patients with severe C. striatum pneumonia showed similar clinical and laboratory features as patients with severe MRSA pneumonia, and both infections were associated with high mortality rates. Further investigations could clarify incidence, clinical characteristics, and outcomes of severe C. striatum pneumonia in critically ill adults and determine whether infections are due to colonization, or community-or healthcare-acquired infections. Clinicians should be aware of this emerging pathogen as a possible cause for severe pneumonia, especially among immunocompromised patients.