PURPOSE: To determine whether radical complete mediastinal lymph node dissection (MLND) improves survival compared with mediastinal lymph node sampling (MLNS) (including non-radical or selective dissection) in patients undergoing resection of non-small cell lung cancer, we performed a meta-analysis of prospective randomized controlled trials and risk-adjusted observational comparative studies.
METHODS: Hazard ratios (HRs) (risk-adjusted HR in case of observational studies) and 95% confidence intervals (CIs) for all-cause death were abstracted from each individual study and combined in both fixed and random-effects models. Meta-regression was conducted for the HR and duration of follow-up or percentage of stage I patients using an unrestricted maximum likelihood model.
RESULTS: Our comprehensive search through March 2011 identified 4 randomized controlled trials and 5 risk-adjusted observational studies. Pooled analysis (3867 patients) demonstrated a statistically significant 25 % reduction in all-cause mortality with MLND relative to MLNS (fixed-effects HR, 0.75; 95% CI, 0.67 to 0.84; P < .00001; P for heterogeneity = .09). When data from randomized trials and risk-adjusted observational studies were pooled separately using a fixed-effects model, MLND was respectively associated with a 22% (HR, 0.78; 95% CI, 0.68 to 0.90; P = .0007) and 31% reduction (HR, 0.69; 95% CI, 0.58 to 0.83; P < .0001) in all-cause mortality relative to MLNS that remained statistically significant. Exclusion of any single study from the analysis did not substantively alter the overall result of our analysis. There was no statistically significant linear relationship between the logarithmic HR and duration of follow-up (P = .59) or percentage of stage I patients (P = .26); and no evidence of significant publication bias (P = .25 by an adjusted rank correlation test; P = .97 by a linear regression test).
CONCLUSIONS: Based on a meta-analysis of 9 studies, MLND may be associated with reduction in all-cause mortality by 25% relative to MLNS in patients undergoing resection of NSCLC.
CLINICAL IMPLICATIONS: To improve survival, MLND rather than MLNS should be considered in patients undergoing resection of NSCLC.
DISCLOSURE: The following authors have nothing to disclose: Masafumi Matsui, Hisato Takagi, Hirotaka Yamamoto, Shin-nosuke Goto, Takuya Umemoto
No Product/Research Disclosure Information