Pneumonia is the leading cause of death in children; the pathogens are often difficult to diagnose. In this study, the performance of metagenomic next-generation sequencing (mNGS) using bronchoalveolar lavage fluid (BALF) samples from 112 children with confirmed pneumonia has been evaluated. mNGS performed a significantly higher positive detection rate (91.07%, 95% confidence interval [CI] 83.80% to 95.40%) and coincidence rate against the final diagnosis (72.32%, 95% CI 62.93% to 80.15%) than that of conventional methods (70.54%, 95% CI 61.06% to 78.58% and 56.25%, 95% CI 46.57% to 65.50%, respectively) (P < 0.01 and P < 0.05, respectively). Bacteria, viruses, and their mixed infections were common in children with pneumonia. Streptococcus pneumoniae was the most common bacterial pathogen in children with pneumonia, while Haemophilus parainfluenzae and Haemophilus influenzae seemed more likely to cause nonsevere pneumonia in children. In contrast, human cytomegalovirus (CMV) infection and the simultaneous bacterial infections could cause severe pneumonia, especially in children with underlying diseases. After adjustments of antibiotics based on mNGS and conventional methods, the conditions improved in 109 (97.32%) children. mNGS of BALF samples has shown great advantages in diagnosing the pathogenic etiology of pneumonia in children, especially when considering the limited volumes of BALF and the previous use of empirical antibiotics, contributing to the timely adjustment of antibiotic treatments, which can potentially improve the prognosis and decrease the mortality. IMPORTANCE Our study indicates high efficiency of mNGS using BALF for the detection of causative pathogens that cause pneumonia in children. mNGS can be a potential diagnostic tool to supplement conventional methods for children's pneumonia.
Keywords: bronchoalveolar lavage fluid (BALF); children; culture; diagnosis; metagenomic next-generation sequencing (mNGS); pneumonia.