Abstract
OBJECTIVE: To compare the efficacy and safety of various pus drainage methods for lung abscess via network meta-analysis. METHODS: Randomized controlled trials (RCTs) from PubMed, Embase, Scopus, Web of Science, VIP, Wanfang, and CNKI were searched up to April 2025. Study quality was assessed using the Cochrane Risk of Bias Tool (v5.4.0). Bayesian network meta-analysis was conducted using Rstudio (v4.4.1). RESULTS: A total of 23 RCTs (1,453 patients) evaluating four drainage techniques were included. Effective rate: CT-guided percutaneous drainage + antibiotics ranked highest (77% probability), outperforming antibiotics alone (OR = 5.4) and conventional postural drainage + antibiotics (OR = 7.4). Hospital stay: Ultrasound-guided drainage + antibiotics ranked best (89% probability), significantly reducing stay versus conventional postural drainage + antibiotics (MD = -19 days). Symptom resolution: Ultrasound-guided drainage + antibiotics ranked highest for cough (SUCRA 97%) and fever resolution (SUCRA 97%). Lesion reduction: Modified postural drainage + antibiotics ranked highest (98% probability), surpassing conventional postural drainage + antibiotics (OR = 0.48) and antibiotics alone (OR = 0.15). CONCLUSION: Comprehensive analysis indicates that ultrasound-guided percutaneous drainage combined with antibiotics demonstrates the highest probability of improvement in key outcome measures, including time to fever resolution, length of hospital stay, and time to cough resolution, potentially showcasing a comprehensive potential for clinical benefit. CT-guided percutaneous drainage combined with conventional antibiotic therapy may be the most effective in terms of the effective rate, while modified postural drainage combined with conventional antibiotic therapy may be the most prominent in increasing the number of lesion reduction. Furthermore, conventional antibiotic therapy alone was inferior to pus drainage procedures combined with conventional antibiotics across multiple outcome measures and was associated with a higher incidence of adverse reactions. However, due to the limited number of available studies, these findings can only be considered preliminary. Future validation through a substantial number of multicenter, large-sample, double-blind randomized controlled trials is warranted. SYSTEMATIC REVIEW REGISTRATION: https://www.crd.york.ac.uk/PROSPERO/, identifier CRD420251127786.