Extent of resection needed to treat lung cancer has long been an issue. The sole randomised controlled trial, reported by the Lung Cancer Study Group, advised against limited resection as standard surgery even for small peripheral non-small-cell lung cancers (< or =3 cm), because of frequent local recurrences. Elsewhere, conflicting results have been reported from different institutions. We therefore conducted a meta-analysis of reported studies to compare survival of stage I patients between limited resection and standard lobectomy. A MEDLINE web search for computer-archived bibliographic data yielded 14 articles suitable for analysis. Combined survival differences (survival rate with lobectomy minus that with limited resection) at 1, 3, and 5 years after resection according to the DerSimonian-Laird random effects model were 0.7% (95% CI, -0.8 to 2.1; P=0.3659), 1.9% (95% CI, -3.7 to 7.4; P=0.5088), and 3.6% (95% CI, -0.4 to 10.5; P=0.3603), respectively. None of these survival differences were significant, indicating that survival after limited resection for stage I lung cancer was comparable to that after lobectomy. However, since interstudy heterogeneity was detected, caution is required in interpretation of the results.