Intrapleural TPA vs Normal Saline in Pediatric Empyema: A Comparative Study

Abstract

OBJECTIVE To determine whether intrapleural tissue plasminogen activator (TPA) reduces hospital length of stay (LOS) vs intrapleural normal saline in children with complicated parapneumonic effusion or empyema. METHODS A retrospective cohort study at a tertiary pediatric referral hospital in Iran, comparing children aged 1 month to 18 years treated during TPA unavailability (2012–2015; intrapleural saline, 40 mL, 1-hour dwell, daily ×3 days) with those treated after TPA became available (2015–2018; TPA 4 mg in 40 mL saline, same protocol). The primary outcome was hospital LOS, and the secondary outcomes were intensive care unit (ICU) stay, fever duration, chest-tube duration, and ICU admission. Multivariable linear regression was adjusted for baseline imbalances between cohorts. RESULTS Of 85 patients, 61 received TPA and 24 saline. The cohorts were comparable at baseline, except for higher pleural fluid adenosine deaminase and shorter time to first instillation in the saline group (medium and very large effects). The mean hospital LOS was 20.28 ± 5.53 days (TPA) vs 19.17 ± 3.60 days (saline). The unadjusted difference was 1.11 days (95% CI, −1.14 to 3.36; P = .366). The adjusted mean difference was 0.36 days (95% CI, −2.91 to 3.63; P = .826). Secondary outcomes showed similarly small differences, with CI ranges compatible with no clinically important effect. CONCLUSIONS Intrapleural TPA showed no detectable benefit over saline in this retrospective comparison. The wide CI ranges indicate insufficient evidence to reject the null hypothesis and not equivalence. Our findings raise, but do not confirm, the hypothesis that mechanical lavage contributes to clinical improvement. Residual confounding and selection bias cannot be excluded. Adequately powered, prospective randomized trials are needed to establish the incremental value of fibrinolysis in children.

The authors' abstract, as published at the source. Hospital Pediatrics, 2026 · DOI ↗