III.

• High-frequency ultrasound and MR were able to detect MPFL lesions after acute lateral patellar dislocation. • High-frequency ultrasound and MR showed similarly high accuracy in diagnosing MPFL lesions. • Interobserver agreement was very good for high-frequency ultrasound and good for MR.

ObjectiveTo evaluate the correlation between the injury patterns of the medial patellofemoral ligament (MPFL) on magnetic resonance imaging in an acute first-time lateral patellar dislocation (LPD) and incidence of a second-time LPD.Materials and MethodsMagnetic resonance images were prospectively analyzed in 147 patients after an acute first-time LPD with identical nonoperative management. The injury patterns of MPFL in acute first-time LPDs were grouped by location and severity for the analysis of the incidence of second-time LPD in a 5-year follow-up. Independent t tests, chi-square tests and Kruskal-Wallis tests were performed as appropriate.ResultsForty-six cases (46/147, 31.3%) of second-time LPD were present at the 5-year follow-up. Fourteen (14/62, 22.6%) and 31 cases (31/80, 38.8%) were present in the partial and complete MPFL tear subgroups, respectively. Twenty-five cases (25/65, 38.5%), 11 cases (11/26, 42.3%), and 8 cases (8/47, 17%) were present in the isolated femoral-side MPFL tear (FEM), combined MPFL tear (COM), and isolated patellar-side MPFL tear (PAT) subgroups, respectively. Compared with the partial MPFL tears, complete tears showed higher incidence of a second-time LPD (p = 0.04). The time interval between the two LPDs was shorter in the complete MPFL tear subgroup (24.2 months) than in the partial tear subgroup (36.9 months, p = 0.001). Compared with the PAT subgroup, the FEM and COM subgroups showed a higher incidence of a second-time LPD (p = 0.025). The time intervals between the two LPDs were shorter in the FEM and COM subgroups (20.8 months and 19.2 months) than in the PAT subgroup (32.5 months, p = 0.049).ConclusionA complete MPFL tear, isolated femoral-side tear and combined tear in a first-time LPD predispose a second-time LPD.