Comparative effectiveness of treatment strategies for pediatric empyema thoracis: a systematic review with pairwise and network meta-analyses

Background Empyema thoracis is a complication of bacterial pneumonia in children with substantial morbidity. Multiple treatment options are available, but the optimal strategy remains uncertain. This systematic review compared major treatment strategies for pediatric empyema thoracis across clinically relevant outcomes. Methods The protocol was prospectively registered with PROSPERO (CRD420251274193). PubMed, Embase, Scopus, Web of Science, and CENTRAL were searched for studies published from January 2000 through January 2026 evaluating empyema thoracis in children aged <18 years. Random-effects pairwise and frequentist network meta-analyses compared chest tube (CT) alone, CT + fibrinolytics, video-assisted thoracoscopic surgery (VATS), and thoracotomy. The primary outcome was hospital length of stay (LOS); secondary outcomes included treatment escalation, treatment success, complications, mortality, and treatment-related durations. Treatments were ranked using P-scores; a sensitivity analysis restricted the network to primary-line arms. Risk of bias was assessed using ROBINS-I, RoB 2, or the JBI critical appraisal checklist, as appropriate to study design, and certainty of evidence was evaluated using GRADE. Results Seventy-seven studies (7,378 children) were included. Compared with CT, treatment escalation was lower with VATS (RR 0.17, 95% CI 0.07–0.37), thoracotomy (RR 0.13, 95% CI 0.04–0.36), and CT + fibrinolytics (RR 0.48, 95% CI 0.26–0.89); certainty was moderate for three of six escalation comparisons and low or very low elsewhere. VATS was associated with shorter pleural drainage (MD −2.54 days, 95% CI −4.64 to −0.44), antibiotic duration (−4.95, −9.36 to −0.54), and fever duration (−2.00, −3.97 to −0.04) but not with LOS (−2.06, −4.57 to 0.45). In pairwise analysis, escalation was higher with CT + fibrinolytics than with VATS (RR 3.18, 95% CI 1.45–6.96). Restricting the network to primary-line arms preserved these associations; the higher network mortality with thoracotomy lost significance. Exploratory mean P-scores were 0.74 (VATS), 0.50 (thoracotomy), 0.47 (CT + fibrinolytics), and 0.29 (CT alone). Conclusions CT alone was associated with more treatment escalation than each alternative, while VATS was associated with shorter treatment-related durations. However, the evidence is predominantly observational, of low to very low certainty for most comparisons, while estimates involving thoracotomy were sensitive to treatment line. These associations are hypothesis-generating and do not establish a treatment hierarchy. Trials with explicit staging and standardized escalation definitions are needed. Systematic Review Registration https://www.crd.york.ac.uk/PROSPERO/view/CRD420251274193 , PROSPERO CRD420251274193.