Abstract Background This systematic review evaluated the evidence comparing patient-important outcomes in spinal or epidural vs general anaesthesia for total hip and total knee arthroplasty. Methods MEDLINE, Ovid EMBASE, EBSCO CINAHL, Thomson Reuters Web of Science, and the Cochrane Central Register of Controlled Trials from inception until March 2015 were searched. Eligible randomized controlled trials or prospective comparative studies investigating mortality, major morbidity, and patient-experience outcomes directly comparing neuraxial (spinal or epidural) with general anaesthesia for total hip arthroplasty, total knee arthroplasty, or both were included. Independent reviewers working in duplicate extracted study characteristics, validity, and outcomes data. Meta-analysis was conducted using the random-effects model. Results We included 29 studies involving 10 488 patients. Compared with general anaesthesia, neuraxial anaesthesia significantly reduced length of stay (weighted mean difference −0.40 days; 95% confidence interval −0.76 to −0.03; P+0.03; I2 73%; 12 studies). No statistically significant differences were found between neuraxial and general anaesthesia for mortality, surgical duration, surgical site or chest infections, nerve palsies, postoperative nausea and vomiting, or thromboembolic disease when antithrombotic prophylaxis was used. Subgroup analyses failed to find statistically significant interactions (P>0.05) based on risk of bias, type of surgery, or type of neuraxial anaesthesia. Conclusion Neuraxial anaesthesia for total hip or total knee arthroplasty, or both appears equally effective without increased morbidity when compared with general anaesthesia. There is limited quantitative evidence to suggest that neuraxial anaesthesia is associated with improved perioperative outcomes. Future investigations should compare intermediate and long-term outcome differences to better inform anaesthesiologists, surgeons, and patients on importance of anaesthetic selection.