A consecutive series of patients undergoing revision total knee arthroplasty was studied prospectively. Clinical and radiographic assessment was performed preoperatively, 6 and 12 months postoperatively, and annually thereafter. Evaluation consisted of a Knee Society clinical score and assessment of patient satisfaction. In addition, patients completed drawings of their lower extremity regarding the location and severity of the pain they experienced preoperatively and at minimum 2-year followup (mean, 36 months; range, 24-48 months). Pain that was localized to the diaphyseal region of the femur or tibia on the drawing was defined as pain at the end of the stem. Clinical, radiographic, and pain drawing data were completed for patients who had 66 of 78 revision total knee arthroplasties performed during the time of the study (85%). All procedures were performed with the same implant system and instrumentation and included fluted cobalt-chrome stems for all patients in whom the stem was implanted without cement and slightly underreamed (press fit). All femoral components had the surface cemented with the stems press fit. Sixteen of the tibial stems were cemented fully, whereas the remaining 50 tibial components were cemented on the surface only with the stems press fit. Localized pain at the end of the stem was present on the femoral side in seven of 66 patients (11%) and in seven of 50 patients with press fit tibial stems (14%). Patients with pain at the end of the stem at 2 to 4 years postoperatively had significantly lower preoperative function scores and overall Knee Society clinical score. Postoperatively, patients with pain at the end of the stem had a significantly lower clinical score; however the postoperative function score and Knee Society clinical score were not significantly different than scores of patients who did not have pain at the end of the stem. There was no correlation between the stem diameter and the occurrence of pain; however, there was a trend for percent canal fill to be higher on the tibial side in patients with pain (71% versus 63%), but this was not statistically significant. Three of the 16 patients with cemented tibial stems (19%) experienced pain at the end of the stem. Patients with press fit stems who had pain at the end of the stem were more likely to express dissatisfaction with the surgical procedure than patients without pain at the end of the stem.