Reproducibility of the semiflexed (metatarsophalangeal) radiographic knee position and automated measurements of medial tibiofemoral joint space width in a multicenter clinical trial of knee osteoarthritis

Objective. To determine the baseline and longitudinal consistency in reproducibility of the semi-flexed metatarsophalangeal (NITP) position in repeat examinations of patients with knee osteoarthritis (OA) recruited for a multicenter clinical trial that terminated within one year (mean duration 0.81 yr), based on precise measurements both of minimum medial tibiofemoral compartment joint space width (JSW) and of tibial inter-rim distance. Methods. Two technologists from 8 and one technologist from 14 clinical radiology units had received previous training in performing nonfluoroscopic semiflexed NITP knee examinations and in quality control criteria for film acceptance. Patients (N = 402; F = 269) were recruited from 58 rheumatology sites and referred to 22 centers, or "x-ray hubs," across North America. At baseline and at study exit, both knees were x-rayed twice on the same day. All films had quality control, and accepted films were digitized at the Central Radiographic Facility and transmitted to the Central Analysis Facility for computerized measurement of minimum medial compartment JSW and tibial inter-rim distance. JSW loss was calculated in the placebo group for the study period. Results. The median SD of the difference in JSW between same-day test/retest film pairs was 0.9 mm for 767 baseline film pairs (knees with JSW > 0 mm), and 0.08 mm for 631 exit film pairs. JSW reproducibility was unaffected by subject's sex, age, and degree of JSW loss. Among all x-ray hubs, JSW reproducibility was excellent in 14 (SD < 0.1 mm), good in 6 (0.1 < SD < 0.2 mm), and moderate in 2 hubs (0.2 < SD < 0.3 mm). No statistical difference was found in technologists' ability either in positioning OA knees or in their test/retest reproducibility in repositioning joints at baseline and at study exit. JSW did not alter significantly during the study period. Conclusion. The protocol for the semiflexed MTP knee position provides a highly reproducible method for anatomically repositioning the knee and for measuring JSW, necessary for OA clinical trials. It is a simple method that can be employed readily at clinical radiology units, as shown by the similarity in JSW precision between x-ray hubs. The results from this large dataset show that throughout the study precise measurements of JSW were obtained from same-day repeat radiographs, findings that together with previous single-center studies confirm the reliability of this method for clinical trial use.