Streptococcus pyogenes Pneumonia in Adults: Clinical Presentation and Molecular Characterization of Isolates 2006-2015

Introduction In the preantibiotic era Streptococcus pyogenes was a common cause of severe pneumonia but currently, except for postinfluenza complications, it is not considered a common cause of community-acquired pneumonia in adults. Aim and Material and Methods This study aimed to identify current clinical episodes of S. pyogenes pneumonia, its relationship with influenza virus circulation and the genotypes of the involved isolates during a decade in a Southern European region (Gipuzkoa, northern Spain). Molecular analysis of isolates included emm, multilocus-sequence typing, and superantigen profile determination. Results Forty episodes were detected (annual incidence 1.1 x 100,000 inhabitants, range 0.29–2.29). Thirty-seven episodes were community-acquired, 21 involved an invasive infection and 10 developed STSS. The associated mortality rate was 20%, with half of the patients dying within 24 hours after admission. Influenza coinfection was confirmed in four patients and suspected in another. The 52.5% of episodes occurred outside the influenza seasonal epidemic. The 67.5% of affected persons were elderly individuals and adults with severe comorbidities, although 13 patients had no comorbidities, 2 of them had a fatal outcome. Eleven clones were identified, the most prevalent being emm1/ST28 (43.6%) causing the most severe cases. Conclusions S. pyogenes pneumonia had a continuous presence frequently unrelated to influenza infection, being rapidly fatal even in previously healthy individuals.


Introduction
The global burden of Streptococcus pyogenes disease is high, causing a wide range of mild to severe clinical manifestations that comprise an important cause of morbidity and mortality worldwide. In high income countries, the way that the disease manifests itself has changed during the last few decades. During the mid-20 th century, S. pyogenes was a common cause of epidemic outbreaks and community-acquired pneumonia (CAP) [1][2][3]. Currently S. pyogenes is considered a rare cause of community acquired pneumonia, being a clinical entity seen only sporadically after an influenza infection [4][5]. However, the involvement of S. pyogenes in lower respiratory tract infections is not infrequently seen during the course of invasive infections, which is associated with an exceptionally high mortality rate [6]. Despite of the severity of the illness, few studies have been designed to comprehensively describe large S. pyogenes pneumonia case series [7,8], with most recent publications being case reports [9][10][11]. Although some molecular and epidemiological data regarding S. pyogenes pneumonic episodes can be extracted from studies reporting invasive S. pyogenes disease, they are not focused on pneumonia and precise data are scarce. The aims of the present study were to describe the clinical features, prognosis and relation with influenza infection of S. pyogenes pneumonia over a decade and to determine the molecular characteristics (emm-type, sequence type, antimicrobial resistance determinants, and superantigen profile) of involved isolates.

Study area and sample
The study was conducted at Hospital Universitario Donostia, which is the referral hospital of the province of Gipuzkoa, northern Spain, and attends a population of about 350,000 adults >18 years old (annual range 348,726 to 354,475 inhabitants), with an annual mean of 50,640 adult admissions. Medical charts of adult patients with a diagnosis of invasive S. pyogenes infection or with a S. pyogenes isolate obtained from lower respiratory tract specimen between January 2006 and December 2015 were revised.
Definition of S. pyogenes pneumonia was based on the presence of clinical findings (cough, fever, sputum production, and pleuritic chest pain), demonstrable infiltrate on chest radiograph or other imaging techniques and a positive S. pyogenes culture obtained from blood, pleural fluid, or bronchial secretions (bronchoalveolar lavage, bronchial aspirate, or sputum). Bronchial secretions yielding S. pyogenes positive cultures were included in the study when it appeared as a single or predominant pathogen. In addition, sputum samples needed to demonstrate > 25 leukocytes and < 10 squamous epithelial cells per low power field on direct Gram-stain.
Pneumonia was considered invasive when S. pyogenes was isolated from a sterile site, or when obtained from a non-sterile lower respiratory tract site but the clinical presentation and analytical data (Streptococcal Toxic Shock Syndrome (STSS), elevated procalcitonin level, etc.) were consistent with invasive disease.
To relate the streptococcal pneumonic cases with the influenza virus circulation, the influenza rate (epidemic threshold 80 cases per 100,000 inhabitants) during the previous weeks and coinciding with each S. pyogenes pneumonia case were recorded (Red Nacional de Vigilancia Epidemiológica. Sistema de Vigilancia de la Gripe en España. ISCIII http://vgripe.isciii.es/ gripe/inicio.do). In patients diagnosed with pneumonia and suspicion of flu during the seasonal influenza period, PCR influenza tests (AH1, AH3 and B) were performed.
Demographic and clinical variables were obtained from the patient's medical charts. Only heavy smokers, those smoking > 20 cigarettes per day, were included among smokers. The Pneumonia Severity Index developed by Fine et al [12] was assigned in each patient. Mortality within 30 days was recorded.

Antimicrobial susceptibility testing
Minimum inhibitory concentrations (MICs) were determined by the broth microdilution method using Sensititre Microtiter Trays (Trek Diagnostics Systems, East Sussex, UK) and cation-adjusted Mueller-Hinton II broth (bioMérieux, Marcy l'Etoile, France) supplemented with 3-5% v/v lysed horse blood. Interpretation was performed according to the Clinical and Laboratory Standards Institute guidelines [13].

Molecular characterization of isolates
All isolates were characterized by sequencing the 180nt of the 5' variable region of the emm gene (http://www.cdc.gov/streplab/protocol-emm-type.html) and by multilocus sequence typing (http://pubmlst.org/spyogenes/). In isolates showing erythromycin resistance, detection of the macrolide resistance genes erm(B), erm(A)TR and mef was performed as previously described [14]. In isolates with reduced susceptibility to fluoroquinolones (ciprofloxacin MIC 2 mg/L) the parC and gyrA genes were sequenced [15]. Clone was defined by the combination of the emm-type, sequence type (ST) and antimicrobial susceptibility pattern.
Detection of ssa, speA, speC and smeZ superantigen genes was performed in two multiplex PCR-s using the chromosomally encoded virulence factor genes speB and slo as successful reaction controls as previously described [16].

Ethics
The study was an observational, laboratory-based, surveillance study with review of medical records. The patient information was anonymized and study investigators and research associates had no direct patient contact and the study protocol involved no change in patient care or management; all decisions regarding patient investigation and treatment were at the discretion of the attending physician. The institutional ethics committee (Ethics Committee for Clinical Research of the Health Area of Gipuzkoa), specifically approved this study.

Statistical Analysis
The associations of the variables of sex, emm-type, STSS, predisposing conditions, bilobar or multilobar pneumonia, season, and patient evolution (death or survival) were calculated by Fisher's exact probability test. To analyze the independent effect of each variable in relation to patient outcome, we performed a logistic regression analysis including death and variables with an initial P value of 0.2 tested in the bivariate analysis. Data were analyzed with the IBM SPSS Statistics, Software version 22.

Incidence
From January 2006 to December 2015, 40 S. pyogenes pneumonia episodes were detected. The number per annual period ranged from 1 to 8 episodes, without any significant trend during the study period (annual average incidence 1.14 episodes per 100,000 inhabitants, range 0.29-2.29).
Most cases occurred in winter (n = 26, 65%) and spring (n = 10, 25%). Nineteen cases (47.5%) occurred when the influenza virus circulation reached the epidemic threshold (Table 1) but only five cases (four confirmed and one suspected) had relation with previous or concomitant influenza infection. There were 8 patients whose pneumonia was diagnosed during a flu period in which a PCR influenza test result was not available. Only one of these 8 patients was suspected of having flu, although in this patient the test was not performed or not recorded.
All but two cases included in the study were hospitalized. The age of patients ranged from 25 to 90 years with a median of 58.4 years and a mean of 60 years. Overall, 32.5% of cases occurred in patients older than 64 years (incidence 1.48 x100,000 elderly people). Nearly twothirds (n = 25, 62.5%) of the patients were male.
All but 3 cases (1 nosocomial and 2 in nursing home residences) were community-acquired. The nosocomial pneumonia in a 66-year-old patient with a kidney transplant and multiple comorbidities, who died 20 days after admission, had abundant growth of S. pyogenes and Pseudomonas aeruginosa in bronchial secretions. This patient had also a cytomegalovirus infection.
Overall, in 20% of cases coinfections were detected, although viral coinfection was not investigated in some cases (Table 1). In addition to the patient coinfected with P. aeruginosa and cytomegalovirus, another 7 coinfections were: 2 with H3N2 Influenza A virus, 1 with H1N1pdm09 influenza A virus, 1 with influenza B virus, 2 with non-encapsulated Haemophilus influenzae, and 1 with Staphylococcus aureus.

Clinical findings
In 21 patients (52.5%), S. pyogenes pneumonia presented as an invasive infection. In 17 of them, S. pyogenes was obtained from blood and in 1 from pleural fluid. The remaining 3 episodes were considered invasive infections even though the microorganism was only isolated in low respiratory secretions, as all had severe sepsis with renal failure, metabolic acidosis, hemodynamic instability, and elevated procalcitonin and C-reactive protein levels. Two of them developed STSS and died.
Twenty-seven patients had severe disease (Fine IV or V). Ten patients developed STSS. In 15 patients, pneumonia was multilobar. In 33 episodes (82.5%), the lower lobes were affected, 11 patients presented with pleural effusion, and cavitation was evidenced in 4 ( Table 1).
Various antibiotic regimens were used, the most commonly prescribed being a combination of a beta-lactam with clindamycin (n = 11) or with levofloxacin (n = 10), monotherapy with a betalactam (n = 10), monotherapy with levofloxacin (n = 7) or a combination of a beta-lactam with clindamycin and levofloxacin (n = 2). No relationship between the treatment received and mortality or clinical course was found.

Antimicrobial Susceptibility
All isolates were susceptible to penicillin (MIC < 0.06 mg/L), and clindamycin (MIC < 0.25mg/L). Only 1 isolate, an emm12/ST36 strain, showed resistance to macrolides; it harbored the mef gene and expressed the M phenotype of resistance (erythromycin MIC = 1 mg/L). Five isolates (12.5%) showed low-level levofloxacin resistance (three isolates MIC = 2 mg/L and two MIC = 4 mg/L), and all were characterized as emm6/ST382 (emm6.0 n = 4 and emm6.4 n = 1), and harbored the Ser79/Ala mutation in the parC gene, but no mutation was detected associated with fluoroquinolone resistance in the gyrA gene.

Discussion
In the 21 st century, S. pyogenes pneumonia has been infrequently reported in detail. Since 2000, only 1 large series has been published describing the clinical presentation, prognosis and characteristics of the isolates [7]. Other publications have reported pneumonia as the clinical manifestation comprising the 7% to 11% of invasive S. pyogenes disease [6,16,17], but did not include specific details of these pneumonic episodes. The former articles, 1 outbreak in a military population [18], and some case reports [9][10][11], constitute the bulk of current information on this clinical entity. Therefore, is not surprising that S. pyogenes is not listed among CAP etiologies [19], or is only mentioned as a sporadic bacterial complication of influenza infection [4]. The incidence of most causes of CAP is not well defined. Apart from S. pneumoniae and Legionella pneumophila, Mycoplasma pneumoniae and Chlamydia pneumoniae are among persistently listed CAP etiologies [4,19]. The pneumonias caused by the latter two bacteria have a relative low fatality rate, and the incidence of pneumonia in adults is not too far from that obtained for S. pyogenes in the present series. In adult and children patients with CAP requiring hospitalization in U.S., Jain et al. found a similar presence of C. pneumoniae as S. pyogenes although a higher prevalence of M. pneumoniae [20,21]. Even so, some of these diagnoses were not specific enough as C. pneumoniae and M. pneumoniae were detected in nasopharyngeal or oropharyngeal swab by means of PCR assay. M. pneumoniae was found to be carried in the upper respiratory tract of a relatively high percentage of healthy, asymptomatic patients [22].
Annual average incidence of S. pyogenes pneumonia was of 1.14 episodes per 100,000 inhabitants (range 0.29-2.29). We assume it represents the minimum incidence as the study was a laboratory based surveillance study and some S. pyogenes pneumonia caseslikely were not included in the study. The main finding in the present series was that S. pyogenes pneumonia was a severe clinical entity with a continuous presence in the community. Most cases occurred in winter and early spring (December to April), as observed in other series [7,8], and mainly affected males, especially elderly or adults with comorbidities. The only source of S. pyogenes infections are human beings and the most frequent focus of infection is the oropharynx. The pathogen can be introduced into the lungs through haematogenous spread or inhalation. The development of pneumonia in some individuals and not in others may depend on both host defenses and the virulence of the microorganism. Among the 40 cases, most patients (26/40 65%) were adults with comorbidities or other important risk factors, although remarkably there were 13 previously healthy people. Six of these healthy people developed severe symptoms (Fine IV or V) and two died.
Among the known risk factors facilitating aspiration of oropharyngeal secretions (decreased consciousness, neurological disorders) surveyed in the study, alcoholism or neurological disorders were evidenced in only 4 patients. Increased age is another risk factor that not only facilitates aspiration of oropharyngeal secretions, but is also associated with a weakened immune response and the presence of more comorbidities. In the present series, half of the patients (n = 20) were older than 60 years. The comorbidities most frequently encountered (hypertension, chronic obstructive pulmonary disease, diabetes mellitus, heart disease, and renal failure) coincided with those commonly found in patients with pneumonia, independently of the causal pathogen [7,23,24]. Apart from mortality, a number of features evidenced the severity of S. pyogenes pneumonia episodes: in more than half of the patients, the pneumonia coursed as an invasive infection, in 18 (45%) patients more than 1 lobe or cavitation was involved, and Fine severity index score placed more than 65% of patients in the intermediate to high risk group.
The 30-day death rate was remarkably high (20%), which elevated to 33.3% among invasive cases and to 30.8% in patients older than 64 years. The fatality rate associated to invasive pneumonia was higher than that associated to other invasive S. pyogenes infections [6,7,16]. Mortality rates higher than 30% have been reported by other authors in pneumonic patients with S. pyogenes invasive infections [7,8]. In the present series half of the patients with a fatal outcome showed rapid onset, dying in less than 24 hours after admission despite the implementation of aggressive supportive care measures and appropriate antibiotic therapy, coinciding with that observed in reports of fulminant S. pyogenes pneumonia [8,9].
The 20% of patients presented with mixed infections, 4 of them (10%) with influenza virus. Most CAP occur in the cold months, coinciding with the circulation of respiratory viruses. The relationship between influenza and S. pyogenes pneumonia in the present series was lower than usually reported [8,[25][26][27]. In the present series only 4 influenza coinfection episodes were confirmed and 1 suspected; and, although 19 pneumonias occurred during periods with high circulation of influenza virus, the remaining 21 episodes occurred outside these periods. The single episode of coinfection with influenza A H1N1pdm09 was detected in 2014, and surprisingly during the course of the main influenza pandemic of 2009, only one S. pyogenes pneumonia case was detected but it was not related with influenza virus. Co-infection with influenza B virus was detected in an elderly woman who died, highlighting the potential morbidity and mortality associated with influenza B virus in the context of concurrent influenza infection [27,28].
All but 3 pneumonic cases were community acquired. There have been reports of clusters of S. pyogenes pneumonia within the same family [29]. However, no clustering with any relationship among episodes was observed in the present study. We found that the same clone (emm1/ ST28) caused an increase of pneumonic episodes in 2012 but no relationship between these patients could be established. This emm1/ST28 clone was the predominant (43.6%) one, followed by emm3/ST15-315 (15.4%) and emm6/ST382 (12.8%) clones. These emm types were also predominant in the series of Muller et al [7]. In 1981-1997, Barnham et al. found that 9 out 17 analyzed pneumonia episodes were caused by an emm1 isolate [8], and the most recently detected S. pyogenes pneumonia outbreak among Marine Corps personnel was due to an emm3/ST15 strain [18].
The increase of emm1/ST28 pneumonia cases each 3-4 years, was perhaps the consequence of some kind of herd immunity after an initial spread of this clone among a susceptible population, although the reason of the dynamics of circulation of clones is unclear.
Although the emm1/ST28 clone in the present series was not associated with death with statistical significance, it was involved in 62.5% of fatal outcomes, including 2 young adults without predisposing factors. Similarly, Santagati et al recently reported 3 cases of fulminant hemorrhagic pneumonia associated with emm1/ST28 strains in previously healthy patients [9]. In the present series, this clone was also related to 11 out of 15 episodes with multilobar pneumonia and in 3 of 4 episodes with cavitary lesions. The virulent potential of emm1 isolates is well documented the world over, with a coincident resurgence of severe invasive infections since the late 1980s [30]. Clinical and epidemiological data analysis, whole-genome sequencing analysis of large bacterial collections, and infection models in nonhuman primates have demonstrated that emm1 strains recovered after 1988 are more virulent than those before 1988 due to acquisition of new genetic material [31].
Fortunately, all isolates were susceptible to clindamycin, which comprises a key antimicrobial for the treatment of severe S. pyogenes infections because it halts exotoxin production. Nonetheless, we detected a worryingly high rate of low level resistance to fluoroquinolones (12.8%) among strains associated with pneumonia, due to the involvement of isolates characterized as emm6/ST382, a clone which intrinsically harbors a mutation in the parC gene [32].
Although the representativeness of the study is limited to a specific geographic area, our findings suggest that today, as in the pre-antibiotic era, S. pyogenes pneumonia remains a very serious clinical entity, especially when it is associated with emm1 strains. S. pyogenes pneumonia is more common than is generally supposed and is often unrelated to influenza infection. Although it affects mainly debilitated people, it can be rapidly fatal even in previously healthy individuals.