BACKGROUND: Varus deformity increases the risk of progression of medial compartment knee osteoarthritis. The aim of this study was to investigate the clinical and radiographic mid-term results of closing-wedge and opening-wedge high tibial osteotomy when used to treat this condition. METHODS: From January 2001 to April 2004, ninety-two patients were randomized to receive either a closing-wedge or an opening-wedge high tibial osteotomy. The clinical outcome and radiographic results were examined preoperatively; at one year; and, for the present study, at six years postoperatively. The outcomes that we reviewed included maintenance of the achieved correction, progression of osteoarthritis (based on the Kellgren and Lawrence classification), severity of pain (as assessed on a visual analog scale [VAS]), knee function (as measured with the Hospital for Special Surgery [HSS] score and Knee injury and Osteoarthritis Outcome Score [KOOS]), walking distance, complications, and survival with conversion to a total knee arthroplasty as the end point. The results were analyzed on the basis of the intention-to-treat principle. RESULTS: Six years postoperatively, the mean hip-knee-ankle (HKA) angle (and standard deviation) was 3.2° ± 4.1° of valgus after a closing-wedge high tibial osteotomy and 1.3° ± 5.0° of valgus after an opening-wedge high tibial osteotomy (p = 0.343). In both groups, the six-year postoperative HKA angles did not differ from the respective one-year postoperative angles. No difference in the severity of pain or in knee function was found between the two groups. Four complications (9%) occurred in the closing-wedge group and seventeen (38%), in the opening-wedge group. Ten (22%) of the patients in the closing-wedge group and three (8%) in the opening-wedge group needed conversion to a total knee arthroplasty within the six-year period (p = 0.05). The difference in the percentage of cases with conversion to total knee arthroplasty was 14% (95% confidence interval [CI] = 21.7 to 0.2). CONCLUSIONS: In the group of patients without conversion to a total knee arthroplasty, there was no difference between the high tibial closing-wedge and opening-wedge osteotomies in terms of clinical outcomes or radiographic alignment at six years postoperatively. Opening-wedge osteotomy was associated with more complications, but closing-wedge osteotomy was associated with more early conversions to total knee arthroplasty. LEVEL OF EVIDENCE: Therapeutic Level I. See Instructions for Authors for a complete description of levels of evidence.