Forty-six studies (n=9,934 patients) were included in the review. Follow-up ranged from 17 to 99 months. Sample sizes ranged from 33 to 929 patients. Thirty-nine studies were of hepatic resection (n=8,485 patients). Seven studies were of radical tumour ablation with or without concomitant resection (n=1,449 patients). The highest-quality study met six of the eight criteria. Four studies met five criteria. Most studies were low quality in terms of definition and measurement of the number of CLMs (reported adequately in six studies), recurrence of the disease (reported adequately in five studies) and all defined confounding factors (reported adequately in 13 studies). Fifteen studies were included in the meta-analysis: 12 that reported hazard ratios and standard errors and three that provided sufficient data from which hazard ratios could be calculated.
Five-year survival (42 papers): For all patients, five-year overall survival ranged from 7% to 58%. Median reported survival was 36% (24 studies). Reported five-year overall survival rates in patients with one to three CLMs ranged from 25.9% to 50%, with a median of 39%. For patients with four or more CLMs, the reported five-year overall survival rate ranged from 0% to 53% (median 17.1%). For patients who underwent resection only, five-year overall survival rate ranged from 25.9% to 50% for those with one to three CLMs (19 studies) and 0% to 53% for those with four or more CLMs (25 studies). For patients who underwent ablation only, reported five-year overall survival rates ranged from 7% to 33% and median overall survival rates ranged from 16% to 35%. Median overall survival in patients with fewer CLMs ranged from 18 to 32 months and ranged from 14 to 34 months in those with more numerous CLMs.
Overall survival was statistically significantly poorer in patients with four or more CLMs compared to those with three or less after undergoing either resection and/or ablation (HR 1.67,95% CI 1.43 to 1.95; n= 15 studies) or resection only (HR 1.74, 95% CI 1.51 to 1.99; n=14 studies). However, there was evidence of moderate statistical heterogeneity for resection and/or ablation only (I2=34. 4%). There was no statistically significant difference in overall survival after ablation only (HR 1.18, 95% CI 0.99 to 1.40; n=one study).
Findings for disease-free survival and median time to relapse were reported in the review.
There was evidence of publication bias for resection and/or ablation.