Abstract
Background
We aimed to investigate the clinical characteristics and influencing factors of severe influenza A virus pneumonia complicated by gram-negative bacterial infection in children.
Methods
We retrospectively analysed data from 43 paediatric patients with severe influenza A virus pneumonia who had been treated at our hospital. The patients were divided into observation (28 patients) and control (15 patients) groups based on whether they had a coinfection with gram-negative bacteria. The clinical data of the two groups were compared, and multivariate regression analysis was performed.
Results
The mean age in the observation group was 2.5 ± 1.53 years; 26 (92.86%) patients were aged ≤5 years, and 11 (39.3%) had underlying diseases. The mean age in the control group was 4.07 ± 2.11 years; 10 (66.67%) patients were aged ≤5 years, and 2 (13.33%) had underlying diseases. The differences in age and the proportion of children with underlying diseases were statistically significant (P < 0.05). The observation group had a longer duration of fever, higher number of patients with fever lasting ≥7 days and higher incidence of gasping, with statistically significant differences (P < 0.05). The C-reactive protein level, procalcitonin level, platelet count and monocyte count in the observation group were significantly higher than those in the control group (P < 0.05). Among the gram-negative bacterial coinfections in the observation group, coinfections with Haemophilus influenzae (21 cases, 75%) and Moraxella catarrhalis (7 cases, 25%) were predominant. The observation group had a higher proportion of paediatric intensive care unit admissions and longer hospital stays than the control group. Additionally, significantly fewer patients in the observation group received neuraminidase inhibitor antiviral therapy within 48 h (P < 0.05). In the observation group, all H. influenzae isolates were sensitive to ceftriaxone, cefotaxime and meropenem, while resistance rates to ampicillin, trimethoprim–sulfamethoxazole, cefuroxime and ampicillin/sulbactam were 60.4%, 75%, 27.8% and 6.2%, respectively. All M. catarrhalis isolates were sensitive to ampicillin/sulbactam, amoxicillin/sulbactam and ciprofloxacin, whereas resistance rates to ampicillin and trimethoprim–sulfamethoxazole were 67.9% and 10.7%, respectively.
Conclusion
Severe influenza A virus pneumonia complicated by gram-negative bacterial infection is more common in children aged <5 years. Younger age and the presence of underlying diseases increase the likelihood of gram-negative bacterial coinfection. Persistent fever, wheezing and significant elevations in C-reactive protein level, procalcitonin level, platelet count and monocyte count are important early warning signs for the clinical identification of gram-negative bacterial infection.
Introduction
Influenza, also known as the flu, is an acute respiratory infectious disease caused by the influenza virus and is capable of triggering global pandemics. During flu seasons, the incidence of influenza in children can reach 20%–30%.1–3 The predominant circulating strains of the influenza virus are influenza A (flu A) and influenza B, with influenza A exhibiting high antigenic variability and posing significant risks. 4 Pneumonia is the most common complication of influenza A in children, and the majority of severe cases and fatalities are attributed to pneumonia.5–7 Secondary bacterial coinfections are frequently occur following influenza virus infection, exacerbating disease severity and affecting prognosis. 8 Our preliminary research found that gram-negative bacteria (GNB) accounted for the majority of severe influenza pneumonia cases with bacterial coinfection in our hospital and were most commonly seen in cases of pneumonia with influenza A virus coinfection. 9 However, few domestic reports have described severe influenza A pneumonia complicated by GNB infection. Therefore, this study further analysed the clinical characteristics and influencing factors of hospitalised children with severe influenza A pneumonia and GNB coinfection to enhance the understanding of this disease and help determine appropriate prevention and treatment strategies for clinicians.
Patients and methods
Patients
This study adopted a retrospective research design and included children with severe influenza A virus pneumonia who were admitted to our hospital between 1 July 2019 and 30 June 2022. All patient details have been de-identified.
The inclusion criteria were as follows:
Aged 1 month to 12 years; Confirmed diagnosis of influenza A virus infection, defined as having an epidemiological history, presenting with influenza-like symptoms and testing positive for influenza A virus antigen; Meeting the diagnostic criteria for severe pneumonia.
Patients who met any one of the following conditions were considered to have severe pneumonia:
Poor general condition; Disturbance of consciousness; Cyanosis; Rapid breathing (age <2 months: respiratory rate (RR) ≥ 60 breaths/min; age ≥2 months to 1 year: RR ≥ 50 breaths/min; age >1–5 years: RR ≥ 40 breaths/min; and age >5 years: RR ≥ 30 breaths/min); Dyspnoea (moaning, nasal flaring and the ‘three depression’ sign, i.e. depression of the suprasternal fossa, supraclavicular fossa and intercostal space); Pulse oxygen saturation ≤92%; Refusal to eat or signs of dehydration; Plain chest radiography or computed tomography (CT) showing lung infiltration, pneumothorax, lung necrosis or lung abscess in ≥two-thirds of a lung; Extrapulmonary complications.
Patients who met the above inclusion criteria were enrolled and divided into two groups based on sputum culture results: those with negative cultures and no signs of bacterial infection were assigned to the control group, and those with GNB infections were included in the observation group. Patients with gram-positive bacterial infections were excluded. The criteria for a positive sputum culture were as follows: tracheal aspirate ≥105 colony-forming units (CFU)/mL or bronchoalveolar lavage fluid (BALF) ≥104 CFU/mL. Bacteria were observed on sputum smear staining, with evidence of leukocyte infiltration and phagocytosis.
The exclusion criteria were as follows:
Incomplete clinical data precluding statistical analysis; Gram-positive bacterial infections identified in cultures (our previous study found that among children with severe influenza–associated pneumonia and bacterial coinfection in our hospital, GNB were the predominant pathogens, while coinfection with gram-positive bacteria was relatively uncommon); Contaminated specimens or colonising bacteria, defined as sputum specimens with >10 squamous epithelial cells per high-power field (HPF) and <25 leukocytes per HPF, which were deemed unqualified; pathogens isolated from nonsterile sites without imaging evidence or clinical signs of infection were considered colonisers.
This study was approved by the Ethics Committee of our hospital (Approval number: XCSER [2023]28). All procedures conformed to the principles of the Declaration of Helsinki. As this was a retrospective analysis, the need for informed consent from patients and their families was exempted. This study followed the relevant Enhancing the QUAlity and Transparency Of health Research (EQUATOR) guidelines. 10
Methods
Data collection
Clinical data of paediatric patients were retrieved from the hospital’s electronic medical record (EMR) system, including clinical data, symptoms, signs, auxiliary examination results, treatment details, prognoses and patient outcomes. The differences between the two groups were compared.
Pathological examination methods
The pathological examination methods were as follows:
Viral detection method. Within 24 h of admission, nasopharyngeal swabs were collected from paediatric patients. Respiratory panel testing for seven viral antigens (including influenza A virus, influenza B virus, adenovirus, respiratory syncytial virus and parainfluenza virus types 1, 2 and 3) was performed using colloidal gold immunochromatography (Standard Diagnostics Inc., South Korea). Additionally, severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) nucleic acid testing was conducted using fluorescent polymerase chain reaction (PCR). Serological detection method. Serum specimens collected from paediatric patients were tested for Mycoplasma pneumoniae antibodies using manual passive agglutination assays. Bacterial detection method. Airway secretions were collected within 24 h of hospitalisation. Nasopharyngeal aspirates were collected using disposable suction, while endotracheal aspirates were obtained via catheter suction under negative pressure in mechanically ventilated patients. For patients undergoing bronchoscopic lavage, BALF was collected in sterile tubes. Sputum specimens showing <10 squamous epithelial cells per HPF on microscopic examination were considered qualified samples, which were subsequently subjected to bacterial culture (Microflex LF/SH, Shanghai, China) and antimicrobial susceptibility testing (VITEK 2 Compact, bioMérieux, France).
Statistical analysis
Data were statistically analysed using SPSS version 23.0 software (IBM Corp., Armonk, NY, USA). Normally distributed measurement data were expressed as mean ± standard deviation (SD) and compared between the two groups using the t-test. Non-normally distributed measurement data were expressed as median (interquartile range) (M (P25, P75)) and compared between the two groups using the Mann–Whitney U test. Count data were expressed as the number (%) of patients and compared between the two groups using the χ2 test. Risk factor analysis was performed using logistic regression analysis, and a P-value <0.05 was considered to indicate statistical significance.
Results
General data
Between 1 July 2019 and 30 June 2022, 49 paediatric cases of severe influenza A virus pneumonia were treated at our hospital. After excluding 6 patients with gram-positive bacterial coinfections, the study included 28 patients with GNB infections in the observation group and 15 patients without bacterial infections in the control group. In the observation group, there were 20 males (71.4%) and 8 females (28.6%), with a mean age of 2.5 ± 1.53 years. Among them, 26 cases (92.9%) were ≤5 years old, while 2 (7.1%) were >5 years old. Among the observed patients, 11 (39.3%) had underlying diseases: febrile seizures (n = 4), congenital heart disease (n = 3), developmental disorders (n = 2), epilepsy (n = 1), asthma (n = 1) and Kawasaki disease (n = 1), with 1 patient presenting with both epilepsy and Kawasaki disease. In the control group, there were nine males (60%) and six females (40%), with a mean age of 4.07 ± 2.11 years. Of these, 13 cases (86.7%) were ≤5 years old, while 2 (13.3%) were >5 years old. The control group included two cases (13.3%) of comorbidities (febrile seizure (n = 1) and growth retardation (n = 1)). Compared with the control group, the observation group had a lower mean age and a higher proportion of children aged ≤5 years with underlying diseases. Demographic comparisons are presented in Table 1.
Comparison of general data between the observation and control groups.
no. (%): number (percentage).
Clinical characteristics
All 28 patients in the observation group presented with high axillary fever, predominantly exhibiting continuous or remittent fever patterns. The median peak temperature reached 39.8°C (range, 39.5°C–40.0°C), with a median febrile duration of 6.5 (range, 5.75–6.5) days. Notably, 15 cases (53.57%) experienced prolonged fever lasting ≥7 days, 21 cases (75%) developed wheezing and 15 cases (53.57%) showed convulsions or consciousness disturbances. Radiological findings revealed pulmonary consolidation with atelectasis in 17 cases (60.7%). One case showed coinfection with respiratory syncytial virus, and three cases were coinfected with Mycoplasma pneumoniae. Severe complications included respiratory failure (two cases), septic shock (one case), toxic encephalopathy (one case), hepatic dysfunction (two cases), metabolic acidosis (one case) and tracheal stenosis (one case). Similarly, all 15 patients in the control group demonstrated high axillary fever, mainly of the continuous or remittent type, with a median peak temperature of 39.0°C (range, 39.4°C–40.2°C) and a shorter median febrile duration of 3.5 (range, 3–6) days. Only two cases (13.3%) had fever lasting ≥7 days. Clinical manifestations included wheezing (five cases, 33.3%), convulsions/consciousness disturbances (four cases, 26.7%) and pulmonary consolidation (five cases, 33.3%). In the control group, none of the children were coinfected with other viruses, and four cases were coinfected with Mycoplasma pneumoniae. Complications were less frequent: respiratory failure (one case), metabolic acidosis (one case), brainstem encephalitis (two cases), hepatic dysfunction (one case), myocardial injury (one case) and agranulocytosis (two cases). Comparative analysis revealed a significantly longer febrile duration in the observation group (P < 0.01), with higher proportions of both prolonged fever lasting ≥7 days (53.57% vs. 13.3%) and wheezing (75% vs. 33.3%) compared with controls. In addition, the observation group showed significantly higher C-reactive protein (CRP) level, procalcitonin (PCT) level, platelet count and monocyte count than the control group. Detailed clinical characteristics are presented in Table 2.
Comparison of clinical characteristics between the observation and control groups.
M (P25, P75): median (interquartile range); no. (%): number (percentage); WBC: white blood cell; NEU: neutrophil; LY: lymphocyte; MO: monocyte; PLT: platelet; CRP: C-reactive protein; PCT: procalcitonin; MP: mycoplasma; ESR: erythrocyte sedimentation rate; CK: creatine kinase; CK-MB: creatine kinase MB isoenzyme; FDP: fibrinogen degradation products; LDH: lactate dehydrogenase.
Bacterial culture and sensitivity
Among the 28 paediatric patients in the observation group, a total of 30 culture-positive specimens were collected, including 28 oropharyngeal/nasal sputum samples and 2 BALF samples. Notably, two cases showed concurrent positivity in both sputum and BALF cultures. The isolated pathogens included Haemophilus influenzae (21 cases, 75%) and Moraxella catarrhalis (7 cases, 25%). All H. influenzae cultures were susceptible to ceftriaxone, cefotaxime and meropenem. The resistance rates to ampicillin, trimethoprim–sulfamethoxazole (TMP–SMX), cefuroxime and ampicillin–sulbactam were 60.4%, 75%, 27.8% and 6.2%, respectively. All M. catarrhalis isolates were susceptible to ampicillin–sulbactam, amoxicillin–sulbactam and ciprofloxacin. The resistance rates to ampicillin and TMP–SMX were 67.9% and 10.7%, respectively.
Treatment and outcomes
Both observation and control groups received neuraminidase inhibitor (NAI) therapy. In the observation group (n = 28), treatment regimens included oral/osmotic nasogastric oseltamivir (26 cases), intravenous peramivir (1 case) and sequential oseltamivir–peramivir therapy (1 case). Notably, only seven cases (25%) initiated NAI treatment within 48 h of symptom onset. All patients in the observation group received antibiotic therapy, with a median duration of 7 (range, 6–8) days. Advanced interventions included fiberoptic bronchoscopy with bronchoalveolar lavage (13 cases, 46.43%) and paediatric intensive care unit (PICU) admission (18 cases, 64.29%). The median hospitalisation duration was 13 (range, 9–15) days, with 22 cases (78.57%) hospitalised for ≥7 days. Mortality occurred in one patient aged ≤5 years with pre-existing epilepsy, who developed hemophagocytic syndrome. In the control group (n = 15), patients received oseltamivir (14 cases) or peramivir (1 case), with 7 cases (46.67%) initiating NAI therapy within 48 h. Fewer patients required bronchoscopy (four cases, 26.67%) or PICU care (two cases, 13.33%). The median hospitalisation duration was shorter at 7 (range, 6–10) days, with eight cases (53.33%) hospitalised for ≥7 days and no mortality. Comparative analysis demonstrated significantly higher rates of PICU admission (64.29% vs. 13.33%, P < 0.01) and prolonged hospitalisation (13 vs. 7 days, P < 0.01) in the observation group. Detailed treatment outcomes are presented in Table 3.
Comparison of treatment between the observation and control groups.
M (P25, P75): median (interquartile range); no. (%): number (percentage); PICU: paediatric intensive care unit.
Risk factors for severe influenza A virus coinfection with GNB
Intergroup comparisons were performed between the observation and control groups (Tables 1 to 3). Variables with statistically significant differences—including age of ≤5 years, underlying diseases and antiviral initiation within ≤48 h—were incorporated into the logistic regression model. The analysis identified age of ≤5 years, antiviral treatment initiation within ≤48 h and underlying diseases as risk factors for severe influenza pneumonia complicated by GNB infection. Specific details are shown in Table 4.
Risk factor analysis of severe influenza A pneumonia complicated by Gram-negative bacterial infection.
SE: standard error; OR: odds ratio; CI: confidence interval.
Discussion
This study provides an in-depth exploration of the etiological characteristics, risk factors and early identification markers of severe influenza A virus pneumonia complicated by GNB infections in children. In recent years, multiple domestic and international reports11–15 as well as our preliminary research 9 have demonstrated a significant shift in the pathogen spectrum of bacterial pneumonia in children—particularly in cases of severe influenza A virus pneumonia with bacterial coinfection—from traditional gram-positive predominance to gram-negative dominance. Our study, conducted in Xiamen, further confirms this epidemiological trend, with GNB accounting for 100% of cases in the observation group and H. influenzae (75%) and M. catarrhalis (25%) identified as the predominant pathogens. The underlying mechanism for this shift may be attributed to the severe damage inflicted on respiratory barriers and systemic immune function by influenza viruses,16,17 creating favourable conditions for the invasion and colonisation of opportunistic gram-negative pathogens such as H. influenzae and M. catarrhalis. This effect is particularly pronounced in young children with immature immune systems.
While elucidating this shift in the pathogen spectrum, the issue of antimicrobial resistance poses significant challenges for clinical management. Our study revealed that H. influenzae exhibits high resistance rates to commonly used antibiotics, including ampicillin (60.4%), TMP–SMX (75%) and cefuroxime (27.8%), consistent with the findings reported by Xiangyu et al. 18 These data strongly suggest that such agents should be avoided as first-line empirical therapy. Fortunately, all H. influenzae isolates in our study maintained high susceptibility to third-generation cephalosporins (ceftriaxone and cefotaxime) and meropenem, providing crucial therapeutic options for severe paediatric cases. These agents should be prioritised as first-line empirical treatment against GNB, particularly for H. influenzae infections. Regarding M. catarrhalis, both our study and the report by Hui et al. 19 confirm its high β-lactamase production rate, resulting in 67.9% resistance to ampicillin. However, β-lactam/β-lactamase inhibitor combinations (such as ampicillin–sulbactam and amoxicillin–clavulanate) demonstrate excellent in vitro activity, making them ideal choices for empirical coverage of M. catarrhalis infections.
Early identification of high-risk populations is crucial for timely intervention. This study clearly delineates two high-risk groups: young children (particularly those <5 years old) and those with underlying medical conditions. In the observation group, 92.86% of children were aged <5 years—significantly higher than that in the control group—and logistic regression analysis confirmed younger age as an independent risk factor for GNB coinfection (odds ratio (OR) = 2.36). This aligns with global data showing that children aged <5 years are at higher risk of severe influenza–related complications and mortality, often associated with bacterial coinfections.20,21 Their vulnerability stems from immature immune systems, making them more susceptible to influenza-induced disruption of immune barriers.15,16 Additionally, the proportion of children with underlying diseases in the observation group (42.86%) was significantly higher than that in the control group (13.33%). Logistic regression analysis identified underlying conditions as an independent risk factor for GNB coinfection (OR = 2.36), consistent with the findings of Yang et al. 20 and studies in Australia and Norway.22,23 Although the result is statistically significant with a high OR, the wide confidence interval (CI) for underlying diseases (95% CI, 1.28–87.2) suggests considerable uncertainty in the estimation. This wide CI may be attributed to the relatively small sample size in this study, which represents a limitation of our research. Despite this uncertainty, it is noteworthy that the lower bound of the CI (1.28) remains greater than 1, indicating that the underlying disease is an important risk factor (OR > 1). This finding alerts clinicians that the underlying condition may represent a highly significant risk factor and should be taken seriously. However, due to the uncertainty in the magnitude of the risk, caution is warranted when applying this information for precise risk assessment in individual patients. Future larger-scale studies—particularly prospective studies or meta-analyses targeting this specific patient population—are needed to narrow the CI and more accurately quantify the true effect size of this risk factor.
Early identification of severe influenza A pneumonia complicated by GNB infection in children is critical for improving clinical outcomes. Our study revealed that compared with children with uncomplicated severe influenza A pneumonia, those with GNB coinfection exhibited more alarming clinical and laboratory features: significantly prolonged fever duration (higher proportion of individuals with fever lasting ≥7 days), increased incidence of wheezing, markedly elevated inflammatory markers (CRP and PCT) and abnormal elevations in monocyte and platelet counts. Notably, disproportionate rises in CRP and PCT serve as sensitive indicators of bacterial infection, particularly GNB infection. The observed thrombocytosis may reflect platelet activation and recruitment during early sepsis,24,25 providing a potential warning sign for impending septic risk. Therefore, for high-risk children with severe influenza A pneumonia (particularly those aged <5 years or those with comorbidities), clinicians should maintain a high index of suspicion for GNB infection when encountering the following conditions: (a) persistent or recurrent high-grade fever; (b) new-onset or worsening wheezing; or (c) concurrent laboratory abnormalities, including significantly elevated CRP/PCT and abnormal platelet or monocyte counts. These indicators should prompt the consideration of empirical GNB coverage even before microbiological confirmation becomes available.
Based on these findings, we propose the following clinical recommendations. First, for children aged <5 years (particularly infants) and those with underlying medical conditions who develop severe influenza A pneumonia, clinicians should maintain a high level of suspicion for GNB infection and implement close monitoring of body temperature, respiratory status and inflammatory markers (CRP and PCT). Second, when children present with either clinical manifestations suggestive of GNB infection (e.g. persistent fever ≥ 7 days or wheezing) or laboratory warning signs (e.g. markedly elevated CRP/PCT levels and abnormal platelet/monocyte counts), empirical anti–gram-negative therapy should be initiated immediately without awaiting definitive microbiological confirmation. Considering the local pathogen distribution profile, empirical therapy should cover both H. influenzae and M. catarrhalis. The first-line agents should be third-generation cephalosporins (e.g. ceftriaxone and cefotaxime). For cases with a high risk of M. catarrhalis infection or when β-lactamase-producing strains are suspected, β-lactam/β-lactamase inhibitor combinations (e.g. amoxicillin/clavulanate and ampicillin/sulbactam) represent appropriate alternatives. Carbapenems (e.g. meropenem) should be reserved for severe cases or suspected multidrug-resistant infections. Concurrently with the initiation of empirical therapy, clinicians must actively and appropriately collect adequate respiratory specimens (e.g. sputum or BALF) for bacterial culture and antimicrobial susceptibility testing. Upon obtaining microbiological identification and susceptibility results, immediate adjustment of the antibiotic regimen is essential to achieve precision therapy (targeted treatment), shorten the treatment duration and minimise the development of resistance.
This study has several limitations. As a single-centre retrospective study, it only included cases from the period between July 2019 and June 2022, with a relatively small sample size (n = 28). The restricted case selection may introduce certain biases, potentially compromising the representativeness of the findings and statistical power. The definition of GNB coinfection was based on sputum and BALF culture results; however, the possibility of contamination or colonisation cannot be ruled out. Furthermore, as all cases were recruited from Xiamen in southern China, regional variations in genetic backgrounds, circulating influenza/bacterial strains and antibiotic prescribing practices may limit the external validity of our results. Future multicentre prospective cohort studies with larger sample sizes encompassing broader geographic regions are needed to more comprehensively characterise the epidemiology, better define risk factors and more accurately determine the prognostic implications of severe influenza A pneumonia with GNB coinfections in Chinese children. Such studies would also enable ongoing surveillance of pathogen distribution shifts and antimicrobial resistance patterns.
In conclusion, this study demonstrates that GNB (predominantly H. influenzae and M. catarrhalis) have emerged as the primary pathogens in paediatric cases of severe influenza A pneumonia with bacterial coinfection in our region. Both age <5 years and underlying comorbidities were independently associated with GNB coinfection in multivariate analysis. The triad of persistent fever (≥7 days), wheezing and marked elevations in inflammatory markers (CRP and PCT) and haematologic indices (platelet and monocyte counts) serves as critical early warning signals for GNB coinfection. In clinical practice, heightened vigilance should be maintained for high-risk paediatric patients through close monitoring of early warning indicators. When GNB coinfection is suspected, empirical antibiotic therapy should be promptly initiated with empirical coverage of both H. influenzae and M. catarrhalis, preferentially using third-generation cephalosporins (e.g. ceftriaxone and cefotaxime) or β-lactam/β-lactamase inhibitor combinations when indicated. Concurrently, comprehensive microbiological investigations should be performed to guide definitive therapeutic decision-making.
Footnotes
Acknowledgements
We thank the patients, nurses and clinical staff who provided care for the patients.
Authors’ contributions
QYW, YL, ZQZ and JJH were involved in study design. QYW and JJH wrote the main manuscript and prepared the tables. FCW, SSY, YPG and YMW facilitated data collection. All authors reviewed the manuscript.
Availability of data and materials
All data generated during this study are included in this published article.
Clinical trial number
Not applicable.
Consent for publication
In this retrospective study, the need for written informed consent from the patients was waived, and the study was approved by the Ethics Committee of Xiamen Children’s Hospital (Approval No. XCSER [2023]28).
Declaration of conflicting interests
The authors declare that they have no competing interests.
Ethics approval and consent to participate
The study was approved by the Ethics Committee of Xiamen Children’s Hospital (Approval No. XCSER [2023]28).
Funding
This work was supported by the Natural Science Foundation of Xiamen Science and Technology Bureau (Grant No.: 3502Z20227147 and 3502Z20227301) and the 1125 Talent Project of Xiamen Children’s Hospital (Grant No.: 2020-1125-YQ-040).
