Discussion
Very few prospective studies were performed for
bacterial/fungal
complications of viral CAP in hospitalized adults. In this present
study, the incidence, outcome, and risk factors of bacterial/fungal
complications were investigated on prospective analysis of two
consecutive years’ data from patients hospitalized with viral pneumonia.
Using comprehensive microbiological tests, our data indicated that
34.5%
of viral CAP-associated hospitalizations had evidence of
bacterial/fungal superinfection. Apart from illness severity and
clinical outcome, the incidence and types of bacterial/fungal infections
with non-influenza respiratory
viruses were also analogous to that associated with influenza,
except
for patients with SARS-CoV-2
infection. The high rate of superinfection complicated the clinical
outcome of patients with viral pneumonia, while patients with
superinfection presented significantly more in-hospital mortality than
those without (22.7% vs 2.4%), especially when they came from non-ICU
wards. In the follow-up period, the incidence of death within 120 days
after admission was significantly higher in patients with
bacterial/fungal superinfection than in those without. Furthermore, we
found that
PaO2/FiO2< 300, BUN ≥ 7.1 mmol/L, leukocytosis, and lymphocytopenia
were independent risk factors of superinfection.
With the development of multiplex PCR assays for viruses, recent
etiology studies have demonstrated an elevated proportion of viral
infection (varies from 6% to 58%) and a declined detection of
bacterial infection (7.8-24.8%)[2, 18, 32]. Flu was the first
ranking viral pathogen, which was in line with other studies [5,
18]. Non-influenza viruses constituted exceeding one-quarter of the
patients with viral pneumonia. Mixed viral infection was diagnosed in
approximately one out of ten patients, indicating that virus
co-infection may be the common phenomenon in viral pneumonia cases, and
multiplexing of molecular assays could facilitate the identification of
those patients [33]. In contrast to secondary bacterial infections,
these mixed viral infections were not associated with a more severe
disease course. Most studies published to date in viral CAP have
included only one influenza season [5, 14]. The present consecutive
2-year prospective study demonstrated a regular seasonality profile in
viral pneumonia, with a peak time in the period of November to March. In
accordance with previous studies, we found comparable clinical
characteristics among influenza, non-influenza, and mixed viral
infection groups [18, 34]. Yet, patients with influenza had a higher
rate of antiviral medication, and so perhaps the shorter length of
hospital stay.
In our cohort, the severity of diseases, complications, and clinical
outcomes of non-influenza pneumonia was analogous to that of influenza
pneumonia, which were in line with previous studies [18, 35].
Accumulating evidence reported the notorious complication of
bacterial/fungal superinfection followed by influenza infection [9],
however, there were few clinical data on whether the risk of
bacterial/fungal superinfection of non-influenza pneumonia was similar
to that of influenza pneumonia. Our data demonstrated that 34.5% of
patients had bacterial/fungal superinfection (ranging from 38%-50% by
different viruses), except for SARS-CoV-2 infection. It was consistent
with reports in previous studies patients with respiratory tract viral
illness [15]. There was limited data on superinfection among
patients with SARS-CoV-2 infection [36]. In our study, 5% of
patients with SARS-CoV-2 had bacterial superinfection, which was
relatively low compared to other respiratory viral infections. The
immune system lacks the preexistence of a clonally expanded population
of antigen-specific lymphocytes during the first year of the SARS-CoV-2
pandemic, the new virus may trigger an exacerbated inflammatory response
that makes superinfection difficult to occur [37]. Notably, there
were no significant differences in the rate of bacterial/fungal
infection by a viral pathogen (χ2
=11.490, P = 0.244). This finding is original, since previous works that
studied bacterial/fungal complications of viral CAP patients did not
make any comparison between influenza and non-influenza viruses.
These findings suggested equal
attention should be given to pneumonia caused by non-influenza viruses.
Influenza infection may facilitate the nasopharyngeal colonization of
opportunistic pathogens (such as A. baumannii , K.
pneumoniae , Pseudomonas species or C. striatum ) that can
gain better access to the lower respiratory tract by chemotaxis and cell
motility genes and cause severe secondary infection [38] [39]
[40]. It has reported that influenza-infected patients co-infected
with P. aeruginosa and Aspergillus spp. were correlated
with a significant mortality [37]. Among antibiotic-resistant
microorganisms from patients in ICU, infection with K. pneumonia ,A. baumannii, and S. aureus was independently associated
with a higher risk of death vs infection with another microorganism
[41-44]. In the present study, the most frequent pathogen identified
in the two years was P.aeruginosa andK. pneumonia , followed byA. baumannii and H. influenza , which was similar to the
trend reported in a multicenter study [14]. Of note, fungal
infection were documented 9 (9/255, 3.5%) patients. Up to 75% ofS. aureus were methicillin-resistant, 53% of K. pneumoniaresistant to β-lactam antibiotics, including third-generation
cephalosporins and carbapenems, 33% of MDR P. aeruginosa and
31% of A. baumannii resistant to carbapenem, which may explain
the higher mortality of patients with superinfection.
In
CAP, viral-bacterial infection has been associated with a more
complicated and worse outcome (e.g., a higher rate of hospital mortality
or mechanical ventilation for > 7 days) than infections
with
viruses
only, bacteria only, or no identified etiology [14, 45]. However,
these conclusions are drawn from patients admitted to the ICU, few
prospective studies were performed for the impact of bacterial/fungal
complications on the overall population of patients with viral CAP. Our
study clearly demonstrated that patients with bacterial/fungal
superinfection had worse clinical outcomes as opposed to those without,
whereas the case-fatality rate was found to be similar between patients
with viral CAP and viral-bacterial CAP in patients from conventional
wards in a recent study [16]. Such a gap probably owed to the older
age and higher frequency of comorbidities of our cohort. Accumulating
experimental data demonstrate that viral infection predisposes to
bacterial superinfection by augmented bacterial adherence and
dysregulation of the immune response [15, 46]. Unexpectedly,
Martin-Loeches et al. [14] found the lack of association between
appropriate antibiotic treatment and the outcome of
viral-bacterial infection, partly
underlining the major role of the immune response in the physiopathology
of patients with severe influenza infection. In fact, respiratory
viruses with both severe and mild disease courses such as influenza,
RSV, and HRV may induce immunosuppression. Type I interferons
productions are shown to increase after respiratory virus infection,
which inhibits the phagocytosis of macrophages, decrease the recruitment
of leukocyte to the lung, and increase levels of anti-inflammatory
cytokines. For adaptive T cells, CD8+T cells are important in recovery
from virulent influenza infections [47]. In our study, we found a
profoundly higher leukocyte count, lower platelet count, and lymphocyte
count. Of note, CD8+ T cells were decreased in the
superinfection group indicating inadequate adaptive immune responses.
These effects all may then contribute to the
susceptibility
toward various superinfections [48].
Given the incidence and worse outcome of superinfection in viral CAP,
some predictors should be identified to determine the initiation of
antimicrobial treatment. In patients with severe influenza pneumonia,
PCT is shown to be a reasonably accurate marker for the detection of
bacterial pneumonia. However, it might not be sufficient as a
stand-alone marker for withholding antibiotic
treatment [49, 50]. More
complementary and simple-to-use markers are needed. Previously, some
data suggest that the superinfection rate increased progressively with
higher admission BUN levels among 12, 363 patients [51]. The current
study confirmed that BUN ≥ 7.1 mmol/L was an important independent
predictor of mixed viral-bacterial/fungal infection. We also found that
hypoxemia might be a good predictor for superinfection, and reflect the
state of alveolar ventilation dysfunction. Further research is needed to
investigate the pathogenesis of hypoxemia in superinfection. Similar to
study in adult patients with respiratory tract viral illness,
leukocytosis was indecently associated with superinfection in patients
with viral CAP. Respiratory virus-induced type I interferons mediate
dysfunction of leukocytes that contributes to the increased
susceptibility of various superinfections [48]. At the same time, we
also found that lymphocytopenia were independent risk factors of
superinfection. Animal model research indicated that lymphocyte
deficiency would deactivate neutrophil and macrophages, resulting in
impaired bacterial clearance [52]. Data from patients with influenza
indicated that lymphopenia on admission was associated with the
occurrence of nosocomial infection, which partially favored our result
[10].
This study has some limitations. First, PCR detection of viruses in the
NP swabs did not indicate the real pathogen of infection. However, low
respiratory tract specimens for viral identification were obtained in
238 (93.3%) in our study, indicated that the causation was believable.
Second, the diagnosis of bacterial infection was largely based on
routine culture methods. The widely empirical use of antibiotics would
result in underestimating the incidence of bacterial superinfection.
Third, the nasopharyngeal microbiota may play a critical role in viral
respiratory infection[39]. Because the
background nasopharyngeal
microbiota of these patients with viral CAP before hospitalization was
unknown, it was hard to trace the isolated bacterial/fungal strains were
due to nosocomial infection or the invading downward of existing
colonizing strains [40]. Longitudinal studies are needed to address
this issue.
In conclusion, superinfections may make up a significant proportion of
total viral CAP cases, and the incidence and types of bacterial/fungal
infections with non-influenza respiratory viruses were also analogous to
that associated with influenza. These findings suggested equal attention
should be given to pneumonia caused by non-influenza viruses. Mixed
viral-bacterial/fungal infections were associated with a higher
mortality rate than primary viral infection. An aggressive
microbiological diagnostic approach should be initiated for those with a
high risk of superinfection, and clinicians should consider the
epidemiology of bacterial pathogens in this setting and consider empiric
treatment for those who are critically ill.