Arthroscopic arthrodesis has been used mainly for in situ fusion of arthritic ankles without deformity. This paper presents the application of arthroscopic arthrodesis of ankles with marked deformity. The results of 78 consecutive cases of arthroscopic ankle arthrodeses, performed in 74 patients, were retrospectively evaluated. Forty-eight ankles had minor deformity (group A), whereas 30 ankles had a varus or valgus deformity of more than 15 degrees (maximum 45 degrees) (group B). The average hospital stay was 3.8 and 3.4 days in groups A and B, respectively (p = 0.74). Postoperative treatment included ankle immobilization for 3 months. Progressive weightbearing was initiated at 2 weeks. Mean followup was 21.1 months. Fusion occurred in 47 of 48 (97.9%) ankles in group A at an average time of 13.1 +/- 5.8 weeks and in 29 of 30 (96.7%) ankles at 11.6 +/- 2.4 weeks in group B (p = 0.19). Unplanned operative procedures were required in 11 ankles (14.1%). One superficial wound infection occurred. Symptomatic arthritis from the adjacent joints developed in six ankles (7.7%). Postoperative ankle alignment in the frontal plane averaged 0.7 and 0.4 degrees of valgus (p = 0.41), whereas the sagittal plane angle averaged 106 +/- 4 degrees and 104.5 +/- 7 degrees in groups A and B, respectively (p = 0.22). The outcome was graded as very good in 79.2% (38 feet) in group A and 80% (24 feet) in group B, fair in 18.8% (9 feet) in group A and 16.7% (5 feet) in group B and poor in one ankle in each group (p = 0.68). The arthroscopic technique offered high fusion rates and low morbidity. Deformity correction was achieved with good results.