The purpose of this study was to report early outcomes after patellar realignment surgery (PRS) was performed for patellar instability in children and adolescents. The study cohort consisted of 23 patients (27 knees) who had PRS between March 1, 2000 and July 30, 2004, by a single surgeon, with a minimum 3-year follow-up and validated outcome measures. Preoperative diagnoses were dislocations in 22 knees and subluxations in 5 knees. The mean age was 14 years and 1 month. Postoperatively, 2 knees had possible recurrent dislocation/subluxation episodes; neither required surgery. Persistent patellofemoral mechanical symptoms have been reported in 10 knees. At a mean follow-up of 5 years and 1 month, the mean Lysholm score was 69.3 and the mean International Knee Documentation Committee (IKDC) score was 65.6. The mean Tegner activity level was 5.4. Patients reported improvement in knee function and pain in 26 of 28 knees (93%). Significant group differences were identified for age at surgery, preoperative diagnosis, and length of preoperative knee symptoms. The group with scores ≥ 70 were younger (13.1 vs. 14.9 y), were all patellar dislocators (vs. 66% in the < 70 group), and with shorter length of symptoms (13 mo vs. 24 mo). In the ≥ 70 groups the Tegner activity levels were higher at 6.8 versus the < 70 group at 4.3. At a mean of 5 years after PRS, 93% of patients reported persistent improvement in knee function and pain, and recurrent patellar dislocation/subluxation episodes were infrequent (7%). Despite the low rate of postoperative patellar instability the patients' subjective opinion of knee function was less than expected (mean IKDC 65.5). Further investigation is warranted to identify causes for this suboptimal outcome at a mean of 5 years postoperatively. These data reinforce the recurring theme documented in other areas of orthopaedics of a disconnection between the "surgical" success (93% improved) and patient's subjective outcome (mean IKDC of 65.6). Level III.