Thirteen studies were included in the review (n=710; range from 20 to 159): 10 studies were single-blind RCTs; two were double-blind RCTs; and one was a prospective blinded case-control study. Most RCTs were of low quality: nine scored 1 on the JADAD scale; two scored 2; and one scored 3. Studies matched on between four and nine criteria to produce well-balanced groups at baseline.
There was no evidence that systemic leukofiltration had an effect on chest tube drainage in the first 24 hours (weighted mean difference was -23.9mL, 95% CI: -95.5 to 47.6mL, p=0.51; 11 studies) or on the total packed red cell blood transfusion requirements (weighted mean difference was 7.84mL, 95% CI: -80.1 to 95.8mL, p=0.85; six studies). There was significant heterogeneity (I2=90.8%) for the chest tube drainage analysis. When the outlying study was removed a statistically significant reduction in chest tube drainage was observed (weighted mean difference was -15.8mL, 95% CI: -30.5 to -1.0mL, p=0.04; 10 studies).
Subgroup analyses showed a statistically significant reduction in chest tube drainage for the systemic leukofiltration group for the six studies that matched on seven or more criteria (weighted mean difference was -18.1mL, 95% CI: -28.3 to -8.0mL, p<0.001) but no evidence of a difference for total packed red cell transfusion. No significant differences were seen for either outcome in those studies with arterial filtration only, or higher-quality studies scoring 2 or more on JADAD. Most of the subgroup analyses reduced the heterogeneity observed in the full analyses.