Field test of the reproducibility of the semiflexed metatarsophalangeal view in repeated radiographic examinations of subjects with osteoarthritis of the knee.

Objective: To estimate the reproducibility of the semiflexed metatarsophalangeal (MTP) view in repeat radiographic examinations of the knee of subjects with osteoarthritis (OA) with respect to radioanatomic alignment of the medial tibial plateau and the central x-ray beam and the precision of measurements of minimum medial tibiofemoral joint space width (JSW).
Methods: Thirty-eight subjects with definite knee OA underwent 2 semiflexed MTP examinations on the same day. Radioanatomic alignment of the medial tibial plateau and the x-ray beam (distance between anterior and posterior margins of the plateau) was measured manually (Lequesne method). Manual measurements of the JSW in repeat radiographs also were obtained. The reproducibility of JSW measurements was estimated by the method of Bland and Altman.
Results: Only 29% of the initial semiflexed MTP radiographs exhibited satisfactory radioanatomic alignment (intermargin distance < or =1 mm). However, intermargin distances in initial and repeat radiographs were highly correlated (r = 0.86, P < 0.01). In 89% of knees, the intermargin distance in the first examination was reproduced (+/- 1 mm) in the second examination. The standard error of measurement (mean of within-knee standard deviations of repeat JSW values) was 0.30 mm. The magnitude of discrepancy between repeat measurements of JSW was related positively to overall radiographic severity of knee OA (P < 0.05).
Conclusion: The semiflexed MTP protocol affords highly reproducible radioanatomic positioning of the knee, although misalignment of the medial tibial plateau and the x-ray beam occurs in >70% of cases. Precision of measurement of JSW in the semiflexed MTP view approaches that associated with fluoroscopically assisted positioning protocols. However, the consequences of poor, albeit reproducible, alignment of the knee in serial semiflexed MTP radiographs in longitudinal studies of OA progression are currently unknown.