Performance of targeted next-generation sequencing in the detection of respiratory pathogens and antimicrobial resistance genes for children
Ruihong Lin,Zhihao Xing,Xiaorong Liu,Qiang Chai,Zefeng Xin,Meng Huang,Chunqing Zhu,Ce Luan,Hongdan Gao,Yao Du,Xuwen Deng,Hetong Zhang,Dongli Ma
DOI: https://doi.org/10.1099/jmm.0.001771
Abstract:Introduction. Respiratory tract infection, which is associated with high morbidity and mortality, occurs frequently in children. At present, the main diagnostic method is culture. However, the low pathogen detection rate of the culture approach prevents timely and accurate diagnosis. Fortunately, next-generation sequencing (NGS) can compensate for the deficiency of culture, and its application in clinical diagnostics has become increasingly available.Gap Statement. Targeted NGS (tNGS) is a platform that can select and enrich specific regions before data enter the NGS pipeline. However, the performance of tNGS in the detection of respiratory pathogens and antimicrobial resistance genes (ARGs) in infections in children is unclear.Aim and methodology. In this study, we estimated the performance of tNGS in the detection of respiratory pathogens and ARGs in 47 bronchoalveolar lavage fluid (BALF) specimens from children using conventional culture and antimicrobial susceptibility testing (AST) as the gold standard.Results. RPIP (Respiratory Pathogen ID/AMR enrichment) sequencing generated almost 500 000 reads for each specimen. In the detection of pathogens, RPIP sequencing showed targeted superiority in detecting difficult-to-culture bacteria, including Mycoplasma pneumoniae. Compared with the results of culture, the sensitivity and specificity of RPIP were 84.4 % (confidence interval 70.5-93.5 %) and 97.7 % (95.9 -98.8%), respectively. Moreover, RPIP results showed that a single infection was detected in 10 of the 47 BALF specimens, and multiple infections were detected in 34, with the largest number of bacterial/viral coinfections. Nevertheless, there were also three specimens where no pathogen was detected. Furthermore, we analysed the drug resistance genes of specimens containing Streptococcus pneumoniae, which was detected in 25 out of 47 specimens in the study. A total of 58 ARGs associated with tetracycline, macrolide-lincosamide-streptogramin, beta-lactams, sulfonamide and aminoglycosides were identified by RPIP in 19 of 25 patients. Using the results of AST as a standard, the coincidence rates of erythromycin, tetracycline, penicillin and sulfonamides were 89.5, 79.0, 36.8 and 42.1 %, respectively.Conclusion. These results demonstrated the superiority of RPIP in pathogen detection, particularly for multiple and difficult-to-culture pathogens, as well as in predicting resistance to erythromycin and tetracycline, which has significance for the accurate diagnosis of pathogenic infection and in the guidance of clinical treatment.