The objective of this systematic review was to compare the outcomes between coronectomy and total removal for third molar extractions with high risk of nerve injury and to help practitioners make prudent decisions on whether and how third molars should be removed. PubMed, Embase, Web of Science, CENTRAL, and SIGLE were searched from January 1990 to October 2011 for randomized or non-randomized controlled trials. Four studies met our inclusion criteria. The pooled risk ratio (coronectomy vs. total removal) was 0.11 (95% CI = 0.03-0.36), 1.03 (95% CI = 0.54-1.98), 0.55 (95% CI = 0.28-1.05), and 1.14 (95% CI = 0.57-2.30) for inferior alveolar nerve injury, post-operative infection, dry socket, and pain at 1 wk after surgery, respectively. A relatively high rate of failed coronectomy in one study (38.3%, compared with 2.3%-9.4% in others) may be attributed to a higher proportion of narrowing roots and vertical impactions. Although root migration rate was high (13.2%-85.29%), the migration distances were short (3.06 ± 1.67 mm), and the directions were away from the nerves. Moreover, the rates of re-operation and root exposure were low. Therefore, coronectomy appears superior to total removal for reducing inferior alveolar nerve damage and could be used in clinical practice for third molar extractions with high risk of nerve injury.