Eight RCTs were included in the review (8,560 participants, range 225 to 3,986). All trials reported adequate sequence generation and allocation concealment and dealt with incomplete outcome data. It was unclear whether trials were blinded or free of selective reporting. Overall the methodological quality of the included trials was considered to be good.
No statistically significant between-group differences were found for overall survival, disease-free survival and regional lymph node recurrence (ipsilateral axilla). There was no evidence of significant statistical heterogeneity. Subgroup analyses did not significantly alter the main findings.
A significantly lower rate of lymphedema six months after surgery (OR 0.24, 95% CI 0.11 to 0.53; five RCTs), numbness or paraesthesia six months after surgery (OR 0.19, 95% CI 0.11 to 0.33; six RCTs), wound infection within the first 30 days (OR 0.50, 95% CI 0.36 to 0.70; two RCTs) and axillary seroma formation within the first 30 days (OR 0.39, 95% CI 0.31 to 0.49; three RCTs) was found in the sentinel lymph node biopsy group compared with the axillary lymph node dissection group. Significant statistical heterogeneity was found for lymphedema (Ι²=85.6%) and paraesthesia (Ι²=92.7%) analyses.
Meta-analysis was not conducted for arm mobility due to mixed data but all included studies found that there was a significantly greater impairment of arm movement (especially arm flexion and abduction) in the axillary lymph node dissection group compared to sentinel lymph node biopsy.
Four of five trials reported that quality of life was better in the sentinel lymph node biopsy group than the axillary lymph node dissection group.
No evidence of publication bias was found. Accuracy results were presented in the paper.