Standing radiographs underestimate joint width: comparison before and after resection of the joint in 34 total knee arthroplasties.

Background: Measurement or estimation of joint width is routinely used in the preoperative evaluation of gonarthrosis. To our knowledge, the validity and reproducibility of this procedure has not been adequately studied.

Patients and methods: We measured joint width in 34 knees (medial arthrosis: n = 22, lateral arthrosis: n = 12) on preoperative weight-bearing radiographs and on radiographs of the corresponding part of the joint after knee arthroplasty. The bone/cartilage pieces were placed in anatomical positions and loaded in a jig made of perspex. High-density film was used to obtain maximum resolution.

Results: In medial and lateral arthrosis, the minimum joint widths were median 0.3 and 0.2 mm smaller on the radiographs of the specimens (p = 0.05, 0.04). In lateral arthrosis the differences were more scattered (95% CI: lateral: 0.1 to -1.2 mm; medial: 0 to -0.5 mm), suggesting less precise determination.

Interpretation: In medial arthrosis, the degree of underestimation is usually small and acceptable. More pronounced discrepancies could be found in lateral arthrosis, calling for the use of further diagnostic measures.