Background. Previous experimental studies have been conducted to evaluate the biomechanical effects of posterior cruciate ligament reconstruction; but no consensus has been reached on the preferred method of reconstruction. Methods. The 3D finite element mesh of a knee joint was reconstructed from computed tomography and magnetic resonance images. The ligaments were considered as hyperelastic materials. The tibiofemoral and patellofemoral joints were modeled with large sliding contact elements. The 3D model was used to simulate knee flexion from 0 degrees to 90 degrees in four cases: a knee with a "native" posterior cruciate ligament, a resected posterior cruciate ligament, a reconstructed single graft posterior cruciate ligament, and a reconstructed double graft posterior cruciate ligament. Findings. A resected posterior cruciate ligament induced high compressive forces in the medial tibiofemoral and patellofemoral compartments. The pressures generated in the tibiofemoral and patellofemoral compartments were nearly the same for the two reconstruction techniques (single graft and double graft). The single graft resulted in lower tensile stresses inside the graft than for the double graft. Interpretation. Firstly, a resected posterior cruciate ligament should be replaced to avoid excessive compressive forces, which are a source of cartilage degeneration. Secondly, the two types of posterior cruciate ligament reconstruction techniques partially restored the biomechanics of the knee in flexion, e.g. contact pressures were restored for pure flexion of the knee. The reconstruction techniques therefore partially restore the biomechanics of the knee in flexion. A double graft reconstruction is subjected to the highest tensile stresses. (c) 2005 Elsevier Ltd. All rights reserved.