| Shoulder surgery | Khan et al., 20241 | Multicentre RCT, 200 patients | Median opioid use halved (56.1 vs 112 mg morphine equivalents, P = .025); better physical function at 2 weeks |
| Total knee arthroplasty | Su et al., 20122 | Multicentre RCT, 11 sites, 187 patients | Lower opioid use in the first 2 weeks (509 vs 680 mg morphine equivalents, P < .05); higher patient satisfaction (P < .0001) |
| Revision total knee arthroplasty | Murgier et al., 20177 | Case-control, 43 patients | Lower total blood loss (260 vs 465 mL, P < .05) and transfusion rate; less pain at rest on day 3 |
| ACL reconstruction | Waterman et al., 20123 | RCT, 36 patients | 83% off all pain medication at 6 weeks vs 28% with ice (P = .0008) |
| ACL reconstruction | Murgier & Cassard, 20144 | Case-control, 39 patients | Less tramadol (57.5 vs 128.6 mg, P = .023), no step 3 analgesics needed; greater knee flexion at discharge (90.5° vs 84.5°, P = .0015) |
| Hip arthroplasty | Leegwater et al., 20125 | Randomised pilot, 26 patients | Smaller post-operative haemoglobin drop and less leg swelling; patients would recommend it to others |
| Hip arthroscopy | Klaber et al., 20196 | Prospective comparative, 40 patients | Lower pain scores than ice (VAS 1 vs 2, P = .0028); 18 of 20 patients satisfied |
| Elective spinal surgery | Nabiyev et al., 20188 | Cohort with historical control, 40 patients | Lower tramadol (63 vs 393 mg, P < .0001), paracetamol and tenoxicam use |
| Paediatric scoliosis surgery | Bellon et al., 20199 | Prospective comparative study | Lower morphine use on days 1 and 3; shorter hospital stay (P = .004) |