Bacterial co-detection in individuals with acute respiratory infections through a direct-to-consumer platform in China: a retrospective cross-sectional study.

BACKGROUND

Although bacterial co-detection is frequently observed in individuals with viral acute respiratory infections, little is known about its epidemiological profile among the broader population in the community. By leveraging a large platform of direct-to-consumer respiratory pathogen testing, we aimed to understand the epidemiological characteristics of bacterial co-detection among individuals with viral acute respiratory infection in the community.

METHODS

Individuals who self-initiated a direct-to-consumer multiplex respiratory pathogen testing service in seven municipalities of China between Nov 1, 2023, and Jan 31, 2025, were included. The primary outcome of interest was co-detection of Streptococcus pneumoniae or Haemophilus influenzae in those testing positive for any of the nine pathogens: influenza A virus (IAV), influenza B virus (IBV), SARS-CoV-2, human adenovirus (HAdV), Mycoplasma pneumoniae, human rhinovirus (HRV), respiratory syncytial virus (RSV), human parainfluenza virus (HPIV), and human coronavirus (HCoV). Both individual-level and population-level risks of bacterial co-detection with each of the nine pathogens were assessed, and subsequently synthesised with random-effects meta-analysis across regions by age group. We further identified clusters of clinical symptoms by the k-means approach and compared the spectrums of symptom clusters between mono-infection and bacterial co-detection groups for each of the nine pathogens.

FINDINGS: 681 436 tests from 645 238 individuals (median age 28 years [IQR 7-38]; 56·0% female) were included in this study. The proportion of bacterial co-detection with S pneumoniae or H influenzae ranged from 75·7% in SARS-CoV-2 to 88·8% in HAdV. Risks of co-detection varied substantially by pathogen. Although increased risk of co-detection with S pneumoniae was observed for HAdV, HRV, IAV, IBV, M pneumoniae, and RSV, decreased risk was observed for SARS-CoV-2 among young children aged five years or younger. For co-detection with H influenzae, increased risk of co-detection was observed in those infected with HAdV, whereas decreased risk of co-detection was observed in those infected with IBV, RSV, and SARS-CoV-2. The risk of co-detection also varied by age group, with children having a higher risk of co-detection with S pneumoniae or H influenzae than adults. The k-means approach identified nine distinct symptom clusters. Different spectrums of symptom clusters were observed between those infected with and without a specific viral pathogen or M pneumoniae, regardless of the bacterial co-detection status. Symptom clusters were generally similar between mono-infection of each of the nine pathogens and co-detection with S pneumoniae or H influenzae, however, there were some exceptions for specific combinations of pathogens and age groups.

INTERPRETATION: Prevalence of bacterial co-detection among individuals infected with respiratory viruses or M pneumoniae is high, particularly among children, and varies by pathogen. Spectrums of clinical symptoms varies by pathogen and age groups, and can be different for the bacterial co-detection group and respective mono-infection group. These findings provide further insights into the possible interplay between viral and bacterial pathogens.

FUNDING: National High Level Hospital Clinical Research Funding; National Natural Science Foundation of China, Noncommunicable Chronic Diseases - National Science and Technology Major Project; Beijing Nova Program.