Objective To compare the discrepancy and consistency in mechanical axis and component position measured by electromagnetic navigation and radiograph in total knee arthroplasty (TKA)to assess whether the navigation system can be used as a substitute for radiograph.Methods A perspective study was performed on 40 cases (61 knees) undergone primary TKA under electromagnetic navigation from July 2006 to December 2006.There were 4 males and 36 females,at a mean age of (66.9 ±8.1) years (range,58-79 years).Mechanical axis angle,distal femoral and proximal tibial cut slope in coronal view (angles cα,β) were recorded both pre-and post-operatively with an intraoperative navigation system and compared against the mechanical axis angle,coronal femoral and tibial slope (angles α,β)measured via full-length radiograph of the lower limb preoperatively and at postoperative 3 months.Consistency in measurement of the same parameters with the two methods was assessed using intraclass coefficiency correlation (ICC).Results Mechanical axis determined by navigation and radiograph showed a mean valgus angle of 9.60° and 9.99° preoperatively and of 1.23° and 1.64° postoperatively,but the two pair parameters revealed no significant differences in the non-parametric test.Mean angle α determined by navigation and radiograph was 89.98° and 88.96° respectively (P ＜ 0.05),and mean angleβ was 90.21 ° and 89.59°respectively (P ＜ 0.05).With deviation value ≤3°,ICC for pre-and post-operative mechanical axis angles,angle α and angleβ was 0.887,0.754,0.632,0.640 respectively.Conclusions Within the acceptable range of deviation,intraoperative navigation data can reflect the pre-and post-operative mechanical axis and prosthesis position evaluated by radiograph.However,the advantages over the consistency of the two measurement methods rest with the evaluation of pre-and post-operative mechanical