RESUMO
BACKGROUND: Proper knee alignment and prosthesis position may theoretically provide better surgical results and increase longevity of total knee arthroplasty. The 3-feet standing long radiograph (LR) is the gold standard for assessment of these parameters. However, the conventional standing regular knee radiograph (RR) is still being used because of convenience and lower cost. We conducted a study to investigate the accuracy of RR compared to LR in assessing the coronal plane prosthesis position. METHODS: We conducted a retrospective cross-sectional study in 100 knee radiographs in 88 patients with knee prostheses. The picture archiving and communications system was used to produce digitized radiographs and perform the angle measurements. LR images were cropped to the same size as the RRs to eradicate rotation error. The femoral component angle (FCA), tibial component angle (TCA), and tibiofemoral angle (TFA) were measured and analyzed by Student t-test. Pearson's correlation coefficient was used to assess interobserver and intraobserver reliability. RESULTS: The RR measurements resulted in a mean increment of 1.3° (95% confidence interval [CI], 0.9°-1.6°; P < .001) for FCA and 1.4° (95% CI, 1.0°-1.9°; P < .001) for TFA compared to LR. The TCAs were similar between 2 techniques (0.2° mean difference; 95% CI, 0°-0.4°; P = .11). RR provided very strong intraobserver reproducibility but only strong interobserver reliability for FCA and TCA while LR provided very strong correlation for all angles. CONCLUSION: RR overestimation of FCA and TFAs in a valgus alignment was minimal, suggesting RR could be an acceptable alternative to LR.