Instability MPFL reconstruction Failure a b s t r a c tPurpose: To ascertain whether differences exist in joint instability after experimentally induced failure of medial patellofemoral ligament (MPFL) reconstruction in the cadaver knee with the four graft types most widely used for this procedure, and whether any of these grafts are associated with decreased risk in the event of failure.Methods: Between March 2011 and March 2012, eight cadavers obtained from the local medical examiner's office were randomly allocated into four groups (four knees each). In each group, a different graft technique was used for MPFL reconstruction. The forces required to induce lateral dislocation of the patella before reconstruction and after experimental failure of surgical reconstruction were recorded. The tested graft techniques were then compared to assess which was associated with the least instability after failed reconstruction.Results: When we compared the groups I (semitendinosus) and II (patellar tendon), the mean differences of the force required to produce a dislocation of the patella before and after the failure were 0.5 N and 12.5 N, respectively (p ¼ 0.028). In comparison between groups I and III (medial third of the quadriceps tendon) the mean differences of the force required to produce dislocation before and after the failure caused were 0.5 N and 22 N, respectively (p < 0.001). In comparison between groups I and IV (Medial third of the quadriceps tendon) we found the mean differences of the force required to produce dislocation before and after the failure caused were 0.5 N and 5 N, respectively (p > 0.999).Conclusions: There were differences in residual instability after simulated MPFL reconstruction failure depending on graft type. Use of the free semitendinosus graft technique was associated with the least risk of residual instability in case of reconstruction failure.