Risk factors for rebound phenomenon (ReP) have not been investigated in detail. Aim of this study is to investigate risk factors of ReP in lower extremity deformities after two-hole plate removal. Patients who underwent hemiepiphysiodes were retrospectively reviewed. Joint orientation angles of tibia and femur, deformity type, age of patient at initial surgery and plate removal, deformity correction rate, bone growth rate were noted. An increase of 5° or more in joint orientation angles in the direction of the initial deformity was considered ReP. A relationship between the patient parameters and ReP was evaluated. Ninety-four segments (55 femurs and 39 tibias) of 43 patients (15 males and 28 females) were included. ReP was observed in 56.2% of all valgus deformities (n = 73), and in 23.8% of varus deformities (n = 21). The ReP rate was 60% (n = 27) and 20% (n = 2) in femoral segments with preoperative valgus and varus deformities, respectively (P = 0.03). When gender, etiology, sides and age of implant application and removal were compared, there was no difference between ReP(+) and ReP(-) groups in femur and tibia. The correction rate of mechanical lateral distal femoral angle in ReP(-) and ReP(+) femurs was 0.7 ± 0.4°/month and 1.1 ± 0.3°/month, respectively (P = 0.01). Age at implant application and removal in the femoral valgus group was significantly lower in the ReP(+) group (P = 0.008 and P = 0.005, respectively). The ReP rate was higher in femoral segments with overcorrection (P = 0.008). The ReP rate was higher in femoral valgus deformities. Younger age at plate application and removal, higher correction rate in deformity and intentional overcorrection increased risk for ReP.