Airway bacterial colonization, biofilms and blooms, and acute respiratory infection

Acute respiratory infections in pediatric patients with tracheostomy are often severe, resulting in prolonged hospitalizations, unscheduled intensive care unit (ICU) admissions and an annual mortality rate of 5% (1–6). Despite this substantive morbidity, mortality and high cost (i.e., estimated at >$1 billion annually (3, 5)), no standardized treatment guidelines exist (7). Indeed, there is much debate in the literature regarding the underlying pathophysiology of acute respiratory tract infections (ARIs) in general (8), let alone ARIs in the population of people with tracheostomy. Recent research into the airway microbiome has shown the potential for colonizing bacteria to enhance susceptibility to and severity of ARIs, particularly those caused by viruses (8–12). The tracheostomized airway in particular offers a unique window into the dynamics of the lower respiratory microbiota that makes it both feasible and ethical to gather the necessary longitudinal samples to understand ARIs. Some of the information from patients with tracheostomy relate to other forms of artificial airway used in the pediatric ICU (PICU) population, but more research is required. In this Concise Clinical Science Review, we discuss 1) bacterial colonization of the airway; 2) blooms of colonizing bacteria during viral infection; 3) colonizing bacteria effect on susceptibility and severity of ARI; and 4) clinical implications including diagnostic approach and future studies.

BACTERIAL COLONIZATION OF THE AIRWAY

During and shortly after birth, the upper airway becomes colonized with a dynamic bacterial flora (13–16) including Staphylococcus aureus, Streptococcus pneumoniae, Haemophilus influenza, and Moraxella catarrhalis, and it is also the primary site of entry for viruses (14, 16). Many factors influence the dynamic composition of the respiratory microbiota in healthy children, including environmental exposures such as antibiotics, diet, and even climate (14, 17–20). Moreover, the pattern of early respiratory tract bacterial colonization affects the stability of the upper airway microbiome later in life (16, 21).

Similar to the native airway, artificial airways such as tracheostomy tubes are rapidly colonized with bacteria shortly after placement, though the mechanisms of colonization and the downstream effects differ (22, 23). Environmental conditions, such as humidified equipment, can support the growth of unique bacteria adherent to the tracheostomy tube (24). The artificial airway also decreases local mechanical host defenses by bypassing the cough reflex and mucociliary function (25). Furthermore, the tube can cause direct mucosal denudation, which fosters a more hospitable environment for bacteria (25). Figure 1, depicting viral ARI pathobiology in children with tracheostomy, demonstrates these conditions on the left, which set the scene for viral-bacterial interactions during ARI, depicted on the right.

Figure 1.

Schema depicting viral ARI pathobiology in children with tracheostomy, which is influenced by, among other factors, airway colonization patterns. Multiple external and internal factors unique to this population contribute to airway colonization. Some factors specific to this population are shown (left). ARI pathophysiology (right) is related to interrelations between the virus, ecological homeostasis (with a proposed bloom of a colonizing bacteria), and host immune response.

Intrinsic factors inherent to microbes such as bacterial virulence factors help circumvent host immune defenses (Figure 1). Some virulence factors faciliate adherent, nutritive biofilms which serve as growth media and as a possible source for immigration into the lower airways (8, 26, 27). These airway bacterial biofilms show a predominance of Staphylococcus aureus, Staphylococcus epidermidis and hypdrophillic bacteria such as Pseudomonas species (28–30). Indeed, a multicenter retrospective cohort study in children with neurologic impairment and ARI, identified Pseudomonas aeruginosa and Staphylococcus aureus as common organisms in lower airway culture samples from bronchoalveolar lavage, endotracheal tubes, sputum, and tracheal aspirate samples (30). Such findings differ from airway studies in otherwise healthy children, which show a dominance of H. influenza and S. pneumoniae (30, 31). Of note, neurologically impaired children with tracheostomy also had higher rates of positive culture results, when compared to those without tracheostomy (30), perhaps reflecting the presence of persistent biofilms.

Clinical factors such as repeated exposure to antibiotics may also lead to changes in the respiratory microbiome and the development of antimicrobial resistance (8, 32). In the ICU, the use of proton pump inhibitors, supine positioning and enteral feeding tubes, can alter the resident microbiome, creating additional disruptions in microbial homeostasis during ARI (8, 33, 34). However, studies specifically characterizing the respiratory microbiome in mechanically ventilated patients are limited in number as compared to those exploring the microbiome of the gut or in chronic lung disease, and the existing studies are difficult to compare given varying methodologies (ie, sampling technique and data analysis) (35). Additional longitudinal, systematic studies are needed to advance the characterization of the respiratory microbiome in mechanically ventilated patients (35).

BLOOMS OF COLONIZING BACTERIA DURING VIRAL INFECTION

Bacterial infections are often temporally associated with preceding viral infection (21, 36–39), and are sometimes referred to as “secondary” infections that are attributed to newly acquired bacterial pathogens. However, studies now suggest that these bacterial co-infections may instead reflect dysbiosis and growth of colonizing bacteria (35). Similar to innate factors associated with the artificial airway, viral presence can create a hospitable environment for colonizing bacteria by inducing changes in the epithelium, decreasing mucociliary clearance and by affecting the host immune response (39) (Figure 1). For decades, clinicians have specifically noted the association between influenza infection and the development of specific bacterial infections, including methicillin resistant Staphylococcus aureus (MRSA) and pneumococcal disease (39–41). A longitudinal study exploring the upper respiratory tract microbiome composition during influenza infection in humans and animal models, found significant disturbances to the “healthy-state” microbiome during viral infection, with normalization of the bacterial community composition as the influenza virus was cleared (42).

The mechanisms underlying these viral induced changes in the resident microbiome are not fully understood, though studies have begun to explore cellular changes in bacteria that may support pathogenic properties. Marks et al demonstrated that cells infected with the influenza A virus and treated with host signals (increased temperature, norepinephrine, extracytoplasmic ATP, and increased nutrient availability) release pneumoccoccus from biofilms, with increased bacterial virulence gene expression that allows for dissemination and invasion (39). Similarly, Smith et al showed respiratory syncytial virus (RSV) infection induced bacterial expression of the virulence factor pneumolysin in a mouse model, supporting the notion that viral infections induce cellular level changes in resident bacteria, thereby increasing disease severity (43). Moreover, studies have shown that in previously healthy, critically ill children admitted with RSV respiratory failure, antibiotic administration within the first 2 days of mechanical ventilation is associated with improved outcomes including faster clinical recovery(44), suggesting the potential for a concomitant bacterial contribution during acute viral infection associated with critical illness (44, 45).

Taken together, these studies invite a clinical reframing, whereby traditional “secondary” bacterial infections may be better thought of as a disturbance of microbial homeostasis (Figure 1). Indeed, several cross sectional studies demonstrate decreased microbiome diversity in patients with pneumonia (46) and microbiome differences in patients admitted with RSV and rhinovirus (47–49). While these studies are limited by the lack of longitudinal data to assess intrasubject variability, further longitudal studies have since demonstrated a similar increase in the relative abundance, or “bloom”, of already present colonizing bacteria in the setting of a viral infection (15, 42, 50–57).

Specifically, Teo et al found that upper airways sampling during the first five years of life, shows a greater abundance of specific Streptococcus, Moraxella, and Haemophilus species during ARI (52). Interestingly, in a study performed in patients with coronavirus disease (COVID-19), the upper respiratory tract microbiota of deceased patients showed a distinct composition when compared to healthy controls and recovered patients, though oropharyngeal microbiota sampling was used as a proxy for the lower respiratory tract, a limitation that was acknowledged in the study (54). A longitudinal study following the composition of nasopharyngeal microbiota in infants throughout the first year of life showed that the relative abundance of Moraxella in the upper airway increased during symptomatic viral infection with or without acute otitis media compared to patients with asymptomatic viral infection and healthy controls (58). Moreover, a Moraxella dominated nasopharyngeal microbiota was associated with upper respiratory infection and sinusitis in pediatric patients followed longtidunially over a year (55). Similarly, a prospective cohort study exploring the microbiota in tracheostomy patients before, during and after ARI found that on day 1 of ARI, the majority of patients had a virus detected and a “bloom” of an existing genera (i.e., Haemophilus and Moraxella) (50). Nonetheless, it remains unclear if these potential blooms represent infections requiring antibiotics, differ by the infecting virus, or are related to ARI severity (49).

COLONIZING BACTERIA EFFECT ON SUSCEPTIBILTY AND SEVERITY OF ARI

Emerging evidence also suggests that the composition of the bacteria colonizing the airway (i.e., microbial homeostasis) has a role in maintaining respiratory health, including susceptibility to respiratory illness (9, 10, 14, 15, 53, 59–62). In children without an artificial airway, colonization with S. pneumoniae, H. influenzae, or M. catarrhalis during early life is associated with increased risk of pneumonia and bronchiolitis, suggesting different colonization patterns may predispose to ARIs (59). Kloepfer et al showed that in the setting of rhinovirus infection, the detection of S. pneumoniae and M. catarrhalis portends increased risk of ARI symptoms and moderate asthma exacerbation (53). Similarly, the nasopharyngeal microbiome in school age children with viral illness and persistent wheeze shifts toward dominance by a small range of pathogenic bacterial genera shortly before detection of viral pathogens or acute onset of symptoms (10). The mechanisms underlying this increased susceptibility are unknown, but focus is turning to colonizing bacteria, and their potential varying effects on illness severity.

Some studies show that patterns of bacterial colonization are associated with increased severity of acute viral illness, particularly in bronchiolitis (17, 63–66). For example, children with a Haemophilus-dominant microbiome profile are at great risk of needing PICU admission for viral bronchiolitis when compared to children with a Moraxella-dominant profile after adjusting for 11 patient-level confounders, including the infecting virus (47). These results have been externally validated in a separate cohort of 307 children (47, 66). Similarly, in a cohort of 132 children age <2 years, De Steenhuijsen Piters et al found that nasaopharyngeal dominance by H. influenzae and Streptococcus were associated with RSV infection and increased ARI severity (i.e., ARI requiring hospitalization) (17). Since these studies are cross sectional, it remains uncertain whether microbial dominance found at the time of hospitalization represents acquisition of new bacteria or blooms of colonizing bacteria. Similarly, recent studies have demonstrated that variation in lung microbiota (bacterial burden and composition) at admission predicts ICU outcomes in mechanically ventilated adult patients (67).

One mechanism underlying these findings may be the association between bacterial colonization patterns and the host response. In influenza A infection, the microbiota influences the generation of CD4 and CD8 T-cell antibody responses, suggesting that the composition of the microbiome is associated with the host response (68, 69). In COVID-19 patients, multiple host factors were significantly correlated with the oropharyngeal microbiota before discharge or death (54). Previous cross-sectional work has identified significant interaction between Haemophilus-dominant microbiota upon hospitalization and systemic (e.g., LL-37) and airway (e.g., CCL5) host responses with regard to the severity of viral ARI (70, 71). In a separate cohort, nasopharyngeal microbiota dominated by H. influenzae or Streptococcus had distinct systemic (i.e., blood) host responses and increased severity of illness as measured by the need for hospitalization (17). Moreover, a recent study in pediatric patients with acute respiratory distress syndrome (ARDS), explored whole blood transcriptomics and found three distinct sub-phenotypes with differing clinical courses and outcomes, though the mechanism underlying these differences are yet to be elucidated (72). Taken together, these studies suggest that the composition of the nasopharyngeal microbiota may be associated with severity by influencing unique systemic and airway host responses (Figure 1).

CLINICAL IMPLICATIONS

Diagnostic approach in suspected ARI

Given the unique bacterial colonization of the artificial airway, and the heightened risk for severe ARI, balancing the need for antibiotics with the risk of antimicrobial resistance is paramount. Determining an informed and systematic approach to using antimicrobial therapy is essential. Currently, the need for antibiotics in the ICU is driven by both clinical (i.e., new physical examination findings, changes in secretion quantity or characteristics, ventilation settings or increased oxygen requirements, chest x-ray imaging findings) (24, 73) and microbiologic criteria. From a microbiologic perspective, clinicians consider tracheal aspirate cultures with gram stain as the gold standard diagnostic test. Quantitiative cultures with ≥ 10⁵ cfu/mL of a single bacterial species are supportive of an ARI requiring antimicrobials, whereas polymicrobial or low levels of bacterial growth support contamination or colonization (74, 75). The presence of microbes not previously identified on prior cultures are also used to support the diagnosis of a new infection, but this measure is unreliable (22, 74–77). The correlation between cultures and the blooming of a bacterial species from longitudinal samples remains unclear.

In regards to Gram stain, the presence of ≥2+ polymorphonuclear cells, or neutrophils, supports active bacterial infection, though this feature too lacks specificity (74) and the presence of neutrophils may not distinguish bacterial colonization from pathogenic process (78). Futhermore, positive identification of a viral infection does not appear to decrease antibiotic use in the PICU given current limitations in accurately diagnosing bacterial respiratory infections (74, 79). Undoubtedly, a deeper understanding of the interplay between colonizing bacteria, pathogenic bacteria and the host response is needed to improve current diagnostic strategies and antimicrobial use.

Future Studies and Clinical Applications

Much remains to be determined in order to develop improved ARI management guidelines and, eventually, novel targets for ARI treatment in those with artificial airways as well as the general population. The traditional clinical approach to ARIs whereby infections are attributed to a single pathogen or category (i.e., viral vs bacterial) ignores the complex interactions of the microbiome with host factors and the environment, particularly in patients with artificial airways, such as in the PICU.

Children with tracheostomy-dependence present a unique need – as well as an opportunity – since they experience frequent infections with resultant repeated exposure to antibiotics, and PICU admission, and are at risk of substantial morbidity and mortality from ARIs (23). How colonizing airway bacteria specifically contribute to ARI susceptibility and severity is unknown in this population (30). Large prospective cohort studies with longitudinal sampling may not only be helpful to tracheostomy-dependent children, but may also inform PICU diagnostic strategies in ARI, decrease unnecessary or incorrect antimicrobial use, and develop novel diagnostic and therapeutic approaches to treating ARIs, including the intentional manipulation of colonizing bacteria (e.g., probiotics). Ultimately, the goal is to understand better ARI pathobiology in all patients.