Establishing the microbial aetiology of pneumonia is crucial in order to provide more targeted therapy, to prevent the overuse of antibiotics (especially broad spectrum antimicrobials) and to avoid complications related to antibiotic use.1 However, despite the implementation of molecular tests and the advances in the diagnosis of pneumonia in the last two decades, the aetiology is identified in only approximately 50% of cases.2,3

In a previous study we found that polymicrobial pneumonia was present in 11% of ICU patients with CAP, and identified it as a risk factor for inappropriate antimicrobial treatment, which is known to increase mortality in these patients.4 We also observed a trend towards higher in-hospital mortality of patients with polymicrobial CAP than in patients with the monomicrobial form. However, the low number of patients in that study meant that it was difficult to present firm conclusions. Recent studies have reported an adverse impact on the severity of the pneumonia and increased readmission and mortality rates in patients with CAP caused by co-infections.5–7 However, little is known about outcomes of polymicrobial pneumonia with regard to the site of care and the microorganism group. The main objective of this study was to investigate the risk factors and outcomes related to polymicrobial aetiology in adult hospitalized patients with CAP.

During the study period, a total of 5114 patients with community-acquired pneumonia (CAP) were admitted to the emergency department. Of these, 916 were excluded for the following reasons: 657 were outpatients, 226 had severe immunosuppression, 45 had missing data, and 18 had active tuberculosis. Therefore, the final study population comprised 4198 patients, of whom 1703 had a microbiological diagnosis and 2495 did not (Fig. 1). Patients included in the study were prospectively evaluated and followed up. Among the 1703 hospitalized patients with CAP with confirmed aetiology, monomicrobial aetiology was identified in 1471 (86%) – 413 ICU patients (83%) and 1058 ward patients (88%) – and polymicrobial diagnoses in 232 (14%) – 84 ICU patients (17%) and 148 ward patients (12%). We compared the outcomes between pneumonias without defined aetiology (2495; 59%) and cases of monomicrobial (1471; 35%) and polymicrobial aetiology (232; 6%). In general, polymicrobial cases had significantly higher rates of ICU admission and in-hospital and 30-day mortality than monomicrobial cases and pneumonias without defined aetiology. Monomicrobial cases had a higher 1-year mortality rate than pneumonia without aetiology, while the 1-year mortality rate of polymicrobial cases was higher than in monomicrobial cases (Supplementary Table 1).

Fig. 1.

Flow chart of the study population.

(0.33MB).

Altogether, 1703 (41%) of the hospitalized CAP patients had a confirmed microbiological aetiology (≥3 microbiological tests performed in 1491 cases, and <3 microbiological tests performed in 212 cases) among whom 1471 patients (86%) presented monomicrobial aetiology (72% admitted to the ward and 28% admitted to the ICU), and 232 patients (14%) had polymicrobial aetiology (64% admitted to the ward and 36% admitted to the ICU) (Fig. 1 and Supplementary Table 2). The specimens obtained included blood cultures from 1360 patients (79%), urinary antigens from 1381 (81%), nasopharyngeal swab from 916 (54%) sputum sample from 890 (52%), acute and convalescent serum sample for serology from 262 (15%), bronchoscopically obtained lower respiratory secretions from 253 (15%), and pleural fluid from 183 (11%). The positive percentages for each test were as follows: blood culture, 12%; pleural fluid culture, 22.3%; bronchial aspirate sample, 53.1%; urinary antigen for S. pneumoniae, 17.1%; urinary antigen for Legionella, 3%; sputum culture, 50%; nasopharyngeal swab, 24%; and serology, 66%.

Median age was 67 years (range 55–81) and most patients were male (61%).

Among 1703 patients with confirmed aetiology, 572 (34%) were treated with β-lactam+macrolide (493 monomicrobial, 79 polymicrobial); 503 (30%) received β-lactam+fluroquinolones (438 monomicrobial, 65 polymicrobial), 298 (17%) received monotherapy with fluoroquinolones (264 monomicrobial, 34 polymicrobial), 106 (6%) received β-lactam monotherapy (90 monomicrobial, 16 polymicrobial) and 177 (10%) received other combinations (144 monomicrobial, 33 polymicrobial).

The main findings of this study of a large population of immunocompetent hospitalized adult patients with CAP are the following: 1. Polymicrobial aetiology was identified in 14% of the overall population; 2. Polymicrobial aetiology was identified in 64% of hospitalized patients in the ward and in 36% of ICU patients; 3. S. pneumoniae and respiratory viruses were the most common pathogens; 4. Polymicrobial pneumonia presented with higher severity and more frequent complications and was associated with higher mortality than monomicrobial pneumonia; 5. Multilobar involvement was an independent predictor of polymicrobial aetiology, while fever was negatively associated with the likelihood of a polymicrobial aetiology; 6. Polymicrobial aetiology was an independent factor associated with 30-day mortality.

Identification of the microbial aetiology of CAP is vital. However, the causative microorganism is overall identified in fewer than 50% of cases, a circumstance that complicates the choice of appropriate empiric antimicrobial therapy. In cases of suspected drug-resistant pathogens, broad-spectrum antibiotics are recommended. In some cases, broad-spectrum antibiotics are used unnecessarily, despite recent evidence of the association between these agents and poor outcomes in CAP patients.12,13 Despite the introduction of molecular tests for the rapid identification of pathogens and resistant genes in recent decades, the lack of standardization and limitation in the interpretation complicates their use in routine clinical practice. Recognizing patients at high risk of polymicrobial infection is important, as it determines whether the initiation of broader-than-usual spectrum empirical antimicrobial therapy is justified, as well as the need for isolation measures to prevent nosocomial spread in some cases.

In a previous study we reported that 11% of patients admitted to the ICU for severe CAP presented polymicrobial pneumonia.4 However, due to the small sample size we were unable to conclude whether the presence of polymicrobial aetiology was related to higher mortality in our study population. In the present study, 14% of our population presented polymicrobial aetiology; most of these cases (64%) were identified among ward patients, while in ICU patients the proportion was lower (36%). Similar results were reported in previous studies which found between 9% and 20% of CAP cases to be polymicrobial, and a similar proportion of polymicrobial cases were admitted to the ICU.5,6,14–17

S. pneumoniae and respiratory viruses were the most frequent pathogens related to pneumonia in our study. It was not a surprise to find that these microorganisms were common in both monomicrobial and polymicrobial pneumonia and in both ward and ICU patients. Other studies, including a recent systemic review, also reported that pneumococcus and respiratory viruses were the most frequent pathogens in CAP.2,3,14,15,18,19

In our study, cases of polymicrobial pneumonia were more severe and more likely to require ICU admission. They presented more complications, such as multilobar involvement, more frequently required invasive mechanical ventilation, had longer ICU and hospital stays, and had higher in-hospital, 30-day, and 1-year mortality. Our findings thus corroborate those of previous studies that reported associations between pneumonia due to co-infection and severe presentations and high mortality rates.5,6,20 Interestingly, patients with negative microbiology had lower rates of in-hospital and 30-day mortality compared to patients with monomicrobial or polymicrobial infections, which corroborates the results of a previous study by our research group that found no association between undiagnosed pneumonia and mortality.21 Polymicrobial aetiology was found to be an independent risk factor of 30-day mortality, alongside age≥65-years, chronic liver disease, chronic neurological disease, confusion, elevated creatinine levels, and ICU admission. This is in accordance with prior reports suggesting that the severity of pneumonia increases exponentially with age and number of comorbidities, which, in turn, increases the risk of death.1,22,23 The severity of the pneumonia, the higher rate of inadequate empiric antibiotic therapy, and the higher rate of complications in polymicrobial infections likely contribute to the high mortality rates, which do not largely differ from prior studies.4,12,24 Interestingly, pneumococcal vaccination and the presence of fever were factors associated with a lower risk of mortality in our study population, possibly because pneumococcal vaccination may reduce the risk of bacterial co-infection caused by the influenza virus and the risk of infection with resistant S. pneumoniae.25 As we reported in our study, the mortality in patients with co-infection was higher than in patients with monomicrobial infection. Our observation that fever was linked to a reduced risk of mortality supports the findings of an earlier study on PES (Pseudomonas aeruginosa, Enterobacteriaceae extended-spectrum β-lactamase–positive, and methicillin-resistant Staphylococcus aureus) pathogens in pneumonia, which also reported that the presence of fever was associated with a lower risk of PES pathogens.26

Multilobar involvement and the absence of fever were the only risk factors for polymicrobial pneumonia identified in our multivariable analyses. These findings are consistent with previous studies, which have reported that polymicrobial infections can worsen outcomes, increase pneumonia severity, and elevate mortality rates in affected patients.20,27,28 For example, the study by Voiriot et al.27 reported that patients with polymicrobial pneumonia had higher rates of a complicated course, including a higher proportion of patients with acute respiratory distress syndrome (ARDS), prolonged mechanical ventilation, and increased hospital mortality. The absence of fever was an unexpected finding, as fever is typically associated with infection. However, in cases of polymicrobial pneumonia, the immune response may be less effective or atypical due to the involvement of multiple pathogens. This absence of fever could reflect a more complex immune interaction, as demonstrated in the study by Schuurman et al.,29 which provided a nuanced assessment of the systemic host response in different CAP etiologies. It should be acknowledged that the identification of these two factors may not be sufficient at this point to justify the initiation of broader-spectrum empirical antimicrobial treatment. Although the multivariable model is well-fitted for polymicrobial aetiology, it has poor discrimination. Further studies are needed to validate and potentially broaden the risk factors that could help identify polymicrobial pneumonia with greater accuracy. Other limitations include the single-centre nature of the study, which may limit the generalizability of the findings; the complete diagnostic workup and microbiological sampling not uniformly applied in all patients given the real-life nature of our study; the potential underestimation of the true incidence of polymicrobial aetiology, as 22% of patients had received prior antibiotic therapy; and our definition of confirmed aetiology in cases of positive PCR for respiratory viruses in the upper airways, which may not be accurate, as we cannot be certain whether the detected virus is the cause of the infection or merely a bystander. This is particularly relevant in our study, as several factors identified as predictors of worse outcomes, such as increasing age and comorbidities, are also associated with delayed clearance of viral infections and increased risks of secondary bacterial infections. Finally, we did not include cases of SARS-CoV-2 to avoid its potential effect on outcomes, particularly during the first and second waves of the pandemic.

In conclusion, polymicrobial CAP is common among hospitalized adult patients. Polymicrobial pneumonia is often severe and associated with high mortality. As S. pneumoniae and respiratory viruses are the main causative agents of polymicrobial pneumonia, strengthening vaccination strategies should be a key approach to addressing this serious form of CAP.

Conceptualization and methodology: CC, DC, AG, MF, MA, AT. Data curation and investigation: CC, AG, AP, DC. Writing—original draft: CC, DC, AG, MA, AP, MF, AT. Writing-review & editing: CC, DC, AT.

This study was supported by CIBER de Enfermedades Respiratorias (CIBERES CB06/06/0028), and by 2009 Support to Research Groups of Catalonia 911, IDIBAPS. The founders of the study had no role in the study design, data collection, analysis, or interpretation, writing of the report, or decision to submit for publication.

The authors declare that they have no competing interest.