Objective: To evaluate the effectiveness of extracorporeal shock wave therapy (ESWT) for the treatment of lateral epicondylitis. Design: Randomized, double blind, placebo-controlled trial. The sample size conferred sufficient power to detect a difference of 20% in outcome scores, and blinding was checked. Setting: Multicenter trial at orthopedic university hospitals, clinics, and practices in Germany and Austria, in the years 1998 to 2000. Participants: Patients were recruited through press reports, health insurance companies, orthopedic practitioners, and hospitals. Eligibility criteria were a diagnosis of radiohumeral epicondylitis on ≥2 of 4 clinical tests, a Roles and Maudsley score of 3 or 4 (range, 1–4; no pain and full activity to pain limiting activity), and ≥6 months of conservative therapy that was terminated ≥2 weeks before entering the trial. Patients with arthritis, neurologic abnormalities, infection, or tumor in the upper extremity to be treated; bilateral symptoms, thrombopathy, or hyperthyroidosis; allergy to mepivacaine or taking anticoagulant therapy; or <18 years of age or pregnant were excluded. Patients (n = 272; mean age, 47 yrs; duration of symptoms, 25 mo; right arm affected, 76%; women, 53%) were assigned to ESWT or placebo (fake ESWT). Intervention: Local anesthetic (3 mL of 1% mepivacaine) preceded 3 weekly treatments of low-energy ESWT (2000 pulses at an energy flux density of 0.07 to 0.09 mJ/mm2) administered by physicians experienced in the technique. The placebo group was assigned to the same regimen, but the shock waves were reflected by means of an air-filled foil. Main outcome measures: Successful treatment was defined as a reduction of the Roles and Maudsley score to 1 or 2 at the 12-week follow-up if the patient had received no additional conservative therapy. Secondary endpoints at 6 and 12 weeks and 12 months were Roles and Maudsley scores, pain, and grip strength. Main results: Data for the 12-week success/failure criterion were available for 90% of the patients. Success rates for the ESWT group and the placebo group were 25.8% and 25.4% (95% CI for the difference, −10.5% to 11.3%). Failure due to Roles and Maudsley scores remaining at 3 or 4 or due to additional treatments received during the period did not differ between groups, or when the data were stratified by center and device used. Roles and Maudsley scores of 1 or 2 did not differ for the groups at 6 or 12 weeks or after 12 months (ESWT, 65.7% vs. placebo, 65.3%; 76% follow-up). Neither pain nor pain-free grip strength differed for the groups at any point. More patients in the ESWT group than the placebo group guessed their group assignment correctly (71% vs. 47%). More minor side effects (principally reddening and swelling of the skin) occurred in the ESWT group. Conclusion: Extracorporeal shock wave therapy was not effective in the treatment of lateral epicondylitis. After 1 year, two thirds of all patients showed improvement.