Epidural analgesia is considered by many to be the best method of pain relief after major surgery. It is used routinely in many thoracic surgery centres. Although effective, side-effects include hypotension, urinary retention, incomplete (or failed) block, and, in rare cases, paraplegia. Paravertebral block (PVB) is an alternative technique that may offer comparable analgesic effectiveness and a better side-effect profile. We undertook a systematic review and meta-analysis of all relevant randomized trials comparing PVB with epidural analgesia in thoracic surgery. Data were abstracted and verified by both authors. Studies were tested for heterogeneity, and meta-analyses were done with random effects or fixed effects models. Weighted mean difference (WMD) was used for numerical outcomes and odds ratio (OR) for dichotomous outcomes, both with 95% CI. We identified 10 trials that had enrolled 520 thoracic surgery patients. All of the trials were small (n<130) and none were blinded. There was no significant difference between PVB and epidural groups for pain scores at 4-8, 24 or 48 h, WMD 0.37 (95% CI: -0.5, 121), 0.05 (-0.6, 0.7), -0.04 (-0.4, 0.3), respectively. Pulmonary complications occurred less often with PVB, OR 0.36 (0.14, 0.92). Urinary retention, OR 0.23 (0.10, 0.51), nausea and vomiting, OR 0.47 (0.24, 0.53), and hypotension, OR 0.23 (0.11, 0.48), were less common with PVB. Rates of failed block were lower in the PVB group, OR 0.28 (0.2, 0.6). PVB and epidural analgesia provide comparable pain relief after thoracic surgery, but PVB has a better side-effect profile and is associated with a reduction in pulmonary complications. PVB can be recommended for major thoracic surgery.