Between November 1985 and June 1987, 751 posterior cruciate-sparing total knee arthroplasties were performed on 523 patients who exhibited fixed varus or valgus deformities. Patients excluded from this study included the following: those with a postoperative follow-up period of less than 2 years (including patients who had died), patients who became infected, and patients with previous failed total knee arthroplasty in the same knee. A total of 473 knees left for evaluation. All arthroplasties were measured using anatomic axis for alignment measurement. The Hospital for Special Surgery scoring system was used to determine the clinical scores prior to the end of each follow-up examination. All ligament releases were performed sequentially, including balancing of the posterior cruciate ligament. All arthroplasties were divided into six separate groups depending upon the degree of varus or valgus deformity. Kaplan-Meier curves were constructed using three methods of failure definition. Curves were then compared between groups. The mean Hospital for Special Surgery score was no different between any of the groups, except for the group of 6 degrees-10 degrees varus, which was significantly higher than the mean score of the 11 degrees and higher valgus group. All other groups were the same statistically. It is concluded that severe varus and valgus deformities may be satisfactorily corrected with the use of a cruciate-retaining type of total knee arthroplasty.