To compare the outcome of 2 arthroscopic techniques for treating recalcitrant lateral epicondylitis. The study included patients undergoing arthroscopic treatment of lateral epicondylitis during 2 different time periods: April 2005 to October 2007 (tenotomy) and May 2009 to June 2010 (debridement). By using a patient-administered form, baseline information including QuickDASH (disabilities of the arm, shoulder and hand) score (primary outcome), visual analog scale (VAS) of pain, and VAS of function was recorded prospectively. To have the same follow-up period of minimum 4 years in the 2 groups, the follow-up was conducted at 2 different points of time. Of a total of 326 patients fulfilling the requirements for inclusion in the study, 283 patients (87%) were followed up (144 male and 139 female, median age 46 [21 to 65] years), 204 (87%) in the tenotomy group and 79 (88%) in the debridement group. In both groups, a significant improvement in the QuickDASH was found at the follow-up compared with baseline: from 60 to 12 in the debridement group (P < .001) and from 59 to 13 in the tenotomy group (P < .001). No statistically significant difference was found in baseline or follow-up QuickDASH, VAS of pain, VAS of function, or failure (reoperation) rate between the 2 groups. The mean length of sick leave was 2 weeks shorter in the debridement only group (P = .007). Both arthroscopic methods lead to a significant improvement of pain and function, and no statistically significant difference was found in any outcome parameters between the 2 techniques at this minimum 4-year evaluation. The results indicate that tenotomy of the extensor carpi radialis brevis may be an unnecessary step in the arthroscopic treatment of lateral epicondylitis Debridement only is a potentially less costly procedure, and the current finding of a mean 2 weeks shorter sick leave in the debridement only group proposes a substantial cost saving in a societal perspective. Level IV, therapeutic case series.