Impacts of environmental factors on the aetiological diagnosis and disease severity of community-acquired pneumonia in China: a multicentre, hospital-based, observational study

Abstract

Environmental exposures are known to be associated with pathogen transmission and immune impairment, but the association of exposures with aetiology and severity of community-acquired pneumonia (CAP) are unclear. A retrospective observational study was conducted at nine hospitals in eight provinces in China from 2014 to 2019. CAP patients were recruited according to inclusion criteria, and respiratory samples were screened for 33 respiratory pathogens using molecular test methods. Sociodemographic, environmental and clinical factors were used to analyze the association with pathogen detection and disease severity by logistic regression models combined with distributed lag nonlinear models. A total of 3323 CAP patients were included, with 709 (21.3%) having severe illness. 2064 (62.1%) patients were positive for at least one pathogen. More severe patients were found in positive group. After adjusting for confounders, particulate matter (PM) 2.5 and 8-h ozone (O₃-8h) were significant association at specific lag periods with detection of influenza viruses and Klebsiella pneumoniae respectively. PM10 and carbon monoxide (CO) showed cumulative effect with severe CAP. Pollutants exposures, especially PM, O₃-8h, and CO should be considered in pathogen detection and severity of CAP to improve the clinical aetiological and disease severity diagnosis.

Keywords: aetiology; community-acquired pneumonia; disease severity; environmental factors; respiratory pathogens.

Graphical abstract

Figure 1.

Flowchart of including patients in the study.

Figure 2.

Pathogen detection in patients with community-acquired pneumonia (CAP). (a) Proportion of detected pathogens in tested CAP patients. (b) Pathogen positivity rate among severe CAP patients. (c) Pathogen codetections in severe (a) and nonsevere (b) CAP patients analyzed by Phi correlation coefficients. C. pneumoniae, Chlamydia pneumoniae; CMV, cytomegalovirus; EV, enterovirus; H. influenzae, Haemophilus influenzae; H. parahaemolyticus, Haemophilus parahaemolyticus; HAdv, human adenovirus; HBoV, human bocavirus; HCoVs, human coronaviruses; HMPVs, human metapneumoviruses; HPIVs, human parainfluenza viruses; HRV, human rhinovirus; IFVs, influenza viruses; K. pneumoniae, Klebsiella pneumoniae; M. catarrhalis, Moraxella catarrhalis; M. pneumoniae, Mycoplasma pneumoniae; P. jirovecii, Pneumocystis jiroveci; RSVs, respiratory syncytial viruses; S. aureus, Staphylococcus aureus; S. pneumoniae, Streptococcus pneumoniae.

Figure 3.

Adjusted ORs (95% CIs) for pathogen detection and severe community-acquired pneumonia (CAP) with increased environmental concentrations according to the logistic regression models. (a) Association of environmental parameters with overall pathogen detection, detection of influenza viruses and Klebsiella pneumoniae, adjusted for age, sex, BMI, temperature, RH, PM2.5, PM10, SO₂, NO₂, O₃-8h, CO, AP, TFSOA, pneumonia severity index score, area, and admission time. (b) Association of environmental parameters with severe CAP in total patients and patients detected with Mycoplasma pneumoniae, adjusted for age, sex, BMI, temperature, RH, PM2.5, PM10, SO₂, NO₂, O₃-8h, CO, AP, TFSOA, area, and admission time. Pathogen detection was extra adjusted in model of total patients. AP, antibiotics pre-admission; BMI, body mass index; CO, carbon monoxide; NO₂, nitrogen dioxide; O₃-8h, 8-h ozone levels; OR, odds ratio; PM, particulate matter; RH, relative humidity; SO₂, sulphur dioxide; TFSOA, time from symptom onset to admission.

Figure 4.

Significant association of specific environmental variables with the detection of specific pathogens and severe community-acquired pneumonia (CAP). (a) For the association of PM2.5 on detection of influenza viruses, exposure-response curve according to distributed lag nonlinear model (DLNM), and single-exposure effects according to Bayesian kernel machine regression (BKMR). The dashed line in DLNM is 75 μg/m³, representing the concentration of emission standard. (b) Exposure-response curve at lag 6 days and single-exposure effects for the association of O₃-8h on detection of Klebsiella pneumoniae. The dashed line is 80 μg/m³, representing half of emission standard. (c) In total CAP patients, exposure-response curve and single-exposure effects for association of PM10 on severe CAP. The dashed line is 75 μg/m³, representing half of emission standard. (d) For the association of CO on severe CAP, exposure-response curve in total CAP patients and single-exposure effects in CAP patients detected with Mycoplasma pneumoniae. The compared concentration of CO is the minimum. Effects from BKMR were defined as the change in the response associated with a change in a particular exposure from its 25th to its 75th percentile, where all of the other exposures are fixed at a specific quantile (0.25, 0.50, or 0.75). CO, carbon monoxide; NO₂, nitrogen dioxide; O₃-8h, 8-h ozone levels; OR, odds ratio; PM, particulate matter; SO₂, sulphur dioxide.