Background Approximately 30–80% of postoperative patients complain about moderate to severe post-surgical pain, indicating that postoperative pain treatment is still a problem. Methods We analysed prospectively collected data on patients in a university hospital receiving systemic and epidural patient-controlled analgesia and continuous peripheral nerve block (CPNB) documented by the acute pain service team in a computer-based system. Results Of 18 925 patients visited in the postoperative period between 1998 and 2006, 14 223 patients received patient-controlled epidural analgesia (PCEA), 1591 i.v. patient-controlled analgesia (IV-PCA), 1737 continuous brachial plexus block, and 1374 continuous femoral/sciatic nerve block. Mean dynamic and resting pain scores (VAS 0–100) were significantly lower for peripheral or neuroaxial regional analgesia compared with patient-controlled systemic opioid analgesia (P<0.05). The risk of a symptomatic spinal mass lesion including epidural haematoma (0.02%; 1:4741) or epidural abscess (0.014%; 1:7142) after PCEA was 1:2857 (0.04%). Neurological complications after CPNB occurred in two patients who received interscalene brachial plexus block. Conclusions We demonstrated that PCEA, IV-PCA, and CPNB are safe and efficient. Although all of these treatment strategies provide effective analgesia, PCEA and CPNB provided superior pain relief compared with IV-PCA. We demonstrated that serious complications of analgesic techniques are rare but possibly disastrous necessitating a close supervision by an acute pain service. We found a low rate of adverse effects including hypotension and motor impairment and a low incidence of epidural haematoma for thoracic PCEA compared with lumbar PCEA.