The coronavirus disease 2019 (COVID-19) outbreak has caused a global health crisis and led to a high rate of critical illness (1). Although the Omicron variant might cause milder cases, mortality increased during the Omicron period, even in a highly vaccinated population (2). Influenza virus infection is also a global public health problem that has caused major morbidity and mortality in many countries (3). Both predispose patients to coinfections and superinfections, especially with bacteria, which could promote severe disease and necessitate timely diagnosis (4–6).

Coinfections in patients with COVID-19 seem uncommon, ranging from 0 to 19% (7–12). However, similar to influenza, superinfections are common in COVID-19, which can follow the initial infection phase or occur during recovery (13). Early recognition of patients with a high risk of coinfections/superinfections is important for the early use of antibiotics and in implementing measures to limit the possibility of superinfection, which may, in turn, reduce mortality, especially in the intensive care unit (ICU). It could even reduce antibiotic resistance caused by the unnecessary use of antibiotics. However, as of March 2023, the risk factors and characteristics related to coinfections/superinfections in COVID-19 and influenza cases in the same ICU have not been described.

Therefore, we aimed to describe the prevalence, pathogenic distribution and clinical characteristics of coinfections/superinfections in patients with COVID-19 and influenza in the same ICU. We also explored the predictive factors of coinfections/superinfections, which might help choose the appropriate application time and variety of antibiotics.

A total of 123 patients with confirmed COVID-19 during the study period were included. The median age of them was 69 (59–78) years, 99/123 (80.5%) of the patients were male. And 87/123 (70.7%) had underlying diseases (60.2% hypertension, 39.8% diabetes, 11.4% chronic heart failure, 22% chronic renal failure, 13.8% chronic lung disease and 40.7% immunocompromised). The Acute Physiology And Chronic Health Evaluation (APACHE) II and Sequential Organ Failure Assessment (SOFA) scores were 17 (12–23) and 6(4–10), respectively. Among them, 78/123 (63.4%) underwent bronchoscopy to obtain bronchoalveolar lavage fluid (BALF) specimens, mNGS of the BALF was performed for 59/123 (48%) patients.

In total, 145 patients with confirmed influenza were identified during three consecutive influenza seasons from December 2017 to February 2022. The median age was 58 (46–69) years, 95/145 (65.5%) were male, and 82 (56.6%) patients had underlying diseases. The APACHE II and SOFA scores were 19 (14–23) and 7 (4–11), respectively (Supplementary Table 3).

In this study, the prevalence of respiratory coinfections/ICU-acquired superinfections in the COVID-19 and influenza cohorts were 33.3%/43.9 and 35.2%/52.4%, respectively. Bacteria were isolated more frequently not only in coinfection but also in ICU-acquired superinfections cases. The most common bacteria identified in coinfection cases were P. aeruginosa, E. faecium, and A. baumannii in patients with COVID-19 and P. aeruginosa, A. baumannii, and K. pneumoniae in patients with influenza. Besides, a significant higher proportion of coinfection events was sustained by Aspergillus spp. The COVID-19 group had more cases of ICU-acquired A. baumannii, C. striatum and K. pneumoniae. P. aeruginosa, A. baumannii, and K. pneumoniae were the three most prevalent pathogens in the influenza cases with ICU-acquired superinfections. In addition, ICU-acquired superinfection was an independent prognostic factor in the COVID-19 group.

Our study demonstrated a higher proportion of patients with coinfections in the COVID-19 group (33.3%). Previous studies and reviews have reported variable coinfection rates, ranging from 3.5 to 14%, focusing on patients from general wards with bacterial infections (7, 8, 11, 12, 22–24). Only a few studies have reported coinfection data from COVID-19 cohort in the ICU (26.9–28%) (11, 25, 26). The coinfection rate in our ICU was higher than that in others. In addition, the main pathogens were gram-negative bacilli and Aspergillus spp., which was not in accordance with previous studies that reported that S. aureus and other common community-acquired bacteria were prevalent in coinfections (8, 11, 26). In our influenza cohort, the prevalence of coinfections was similar to most of previous study (4, 27), but higher than others (28, 29).

Meanwhile, unlike other studies, which reported S. aureus, S. pneumoniae and Hemophilus influenza were the most commonly isolated co-infectious agents (4, 5, 27, 29), we found typical pathogens of nosocomial infections such as A. baumannii and P. aeruginosa. Some factors may explain this phenomenon. First, in our study, the time from onset to ICU admission was longer than in other studies (9 vs. 3–5.6 days). Only, 17.2% of the patients in our study were admitted directly from the emergency or outpatient department. The rest were hospitalized in general wards and other ICUs before being transferred to the ICU in our study. Second, we supposed that it was due to the inclusion bias in different studies but not the actual situation. The diagnostic criteria for coinfection were not standardized, which confounded pathogenicity and colonization. Most of studies that assumed coinfection as the pathogen were detected. In addition, the time of diagnosis of coinfection was not uniform. Many studies included secondary infection or mixed coinfection and secondary infection. Obviously, the proportion of coinfection with Aspergillus spp. was quite high in our study. Awareness of the possibility of fungal coinfection in COVID-19 is essential to initiate empirical antifungal therapy and fungal infection test as early as possible, which assisted in preventing severe illness and death from coinfection.

In the COVID-19 cohort, patients aged 50–70 years had the highest prevalence of coinfection (OR: 2.680), which was in agreement with studies that reported that older adult patients tended to have coinfection (30, 31). Hughes et al. (8) indicated that the age group of 55–81 years were predisposed to coinfection, which concorded with our study (50–70 years). APACHE II ≥18 and CD8+ T cell ≤90/μL were also independent risk factors for coinfection. These findings might imply that critical conditions due to disease and declined immune ability due to aging are the causes of coinfection in the older adult (32). In addition, BMI ≤23.5 kg/m² and WBC ≥10 × 10⁹/L were predictive factors of influenza coinfection. A higher WBC count was a manifestation of coinfection. A large prospective study conducted by Langouche et al. reported that critical illness evokes adipose tissue accumulation of alternatively active M2 macrophages, which have local anti-inflammatory functions (33). Patients with a lower BMI may have weak anti-inflammatory abilities. Our study showed that patients with high-risk factors could be treated empirically with antibiotics after ICU admission. Antibiotics should cover both gram-positive and gram-negative bacteria. It should also be noted that patients with COVID-19 may have coinfection with fungi, and these patients should be promptly treated with antifungal therapy.

The pathogenesis of influenza coinfection has been elaborated. Influenza virus contributes to respiratory epithelial cell damage, bacterial mucociliary clearance dysfunction, and immune response dampening, enabling increased bacterial adherence and invasion (34, 35). As for COVID-19, the mechanism of the pathogenesis of coinfection remains indistinct, and we lack evidence to support the bacteria-virus association (7).

The incidence of ICU-acquired superinfections was similar in the COVID-19 and influenza cohort (43.9% vs. 53.4%). In both cohorts, gram-negative bacilli were responsible for most ICU-acquired superinfections. A. baumannii, K. pneumoniae, and P. aeruginosa were the most commonly identified bacteria. However, the COVID-19 group had significantly higher rates of C. striatum and lower rates of P. aeruginosa, S. maltophilia, B. cenocepacia, and CMV than the influenza cohort. Possibly due to different periods, the distribution of nosocomial bacteria in the ICU was different. CMV infection is generally asymptomatic and usually presents as a latent infection in healthy individuals. Immune abnormalities caused by influenza may promote CMV expression (36).

Further, the incidence of MDR bacteria was similar between COVID-19 and influenza. In the COVID-19 group, 2 patients had ICU-acquired superinfections with Aspergillus spp. and Rhizopus spp. They all experienced kidney transplantation and took immunosuppressants for long periods, which accounted for the immune disorder. In a multicenter cohort study conducted by Rouze et al. (37), the incidence of ventilator-associated lower respiratory tract infection was significantly higher in the COVID-19 group (50.5%) than in the influenza group (30.3%). However, we did not reach this conclusion. The shorter duration of IPPV for COVID-19 in our study may have contributed to this difference. Like VAP in other diseases (38), ICU-acquired superinfection was associated with longer IPPV time, ICU stay and hospital stay. Another key finding from our study was that ICU-acquired superinfection was associated with a reduced survival rate in patients with COVID-19. This results agrees with those of other studies that have shown a negative association between secondary infection and an increased risk of death (39, 40).

This study identified many predictive factors for ICU-acquired superinfection of COVID-19. The decrease in CD8+ T cells and the increase of IL-8 levels indicated the cytokine storm activation and subsequent immunosuppression (41). Immune response dysfunction may be associated with a higher risk of ICU-acquired superinfection. Moreover, high blood glucose levels, hypertension, and smoking have all reported to be related to secondary infection (42–44). Based on the above results, early identification of high-risk patients for ICU-acquired superinfection and active examination for pathogens will facilitate timely and appropriate antibiotics, which is beneficial to prognosis.

To our knowledge, this is the first study to focus on the coinfections/ICU-acquired superinfections of COVID-19 versus influenza ICU patients in Asia. We clearly defined the time of sample positivity time from the date of ICU admission (<48 h vs. ≥ 48 h). However, this study had several limitations. First, this was a single-center retrospective study, and the study population was relatively small; therefore, possible selection and report biases exist, and it is difficult to generalize the results to other centers. Second, ICU management, ICU isolation measures, sampling methods, and infection diagnostic techniques differed between the COVID-19 and influenza pandemics. In addition, we distinguished infection and colonization using clinical judgment rather than bacterial count, which may have affected the detection rate. Most importantly, some patients were exposed to antibiotics and stayed in general wards or other ICUs before our ICU admission. This could have impacted pathogenic microorganism detection and potentially underestimated or overestimated the real coinfection rate.

Coinfections and ICU-acquired superinfections were frequent not only in COVID-19 patients but also in influenza patients admitted to the ICU. The represent agents of coinfections in ICU patients were different from those in the general ward. Our study provides evidence supporting close monitoring and empirical choice of antibiotics according to the pathogen for COVID-19 and influenza cases at risk of coinfections/ICU-acquired superinfections in the ICU. Apart from the limited study population, ICU management, ICU isolation measures, sampling methods, and infection diagnostic techniques may have impact on our conclusion. A large-scale and well-designed RCT is needed in the future.

The original contributions presented in the study are included in the article/Supplementary material, further inquiries can be directed to the corresponding authors.

The studies involving humans were approved by China-Japan Friendship Hospital. The studies were conducted in accordance with the local legislation and institutional requirements. The ethics committee/institutional review board waived the requirement of written informed consent for participation from the participants or the participants’ legal guardians/next of kin due to the study’s retrospective nature.

LH and QZ took full responsibility for the integrity of the submission and publication and were involved in the study design. QZ, LH, and ZC participated in the design of the study and coordination. ZC involved in data collection, had full access to all of the data in the study, took responsibility for the integrity of the data, and were responsible for data verification. ZC and LH took the responsibility for statistical analysis and drafted the manuscript. QZ provided crucial revision for important intellectual content. All authors made substantial contributions to the conception and design of the study or to the data acquisition, analysis or interpretation, reviewed and approved the final manuscript, and significantly contributed to this study.

The funding source of the study [CAMS Innovation Fund for Medical Sciences (CIFMS) 2022-I2M-JB-016] is an academic non-profit organization that played no role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.