STANDING KNEE RADIOGRAPHS: AP VS PA PROSPECTIVE EVALUATION ARTHROSCOPICALLY CONFIRMED ARTHRITIS

Kurt P. Spindler, MD, Robert Boyce, MD, Eric C. McCarty, MD, Todd Michener, BS, and Yu Shyr, MD

Objective:
Despite the frequency of radiographs (XR) used for screening and diagnostic purposes to detect knee arthritis, no prospective studies have investigated the routine use in general sports medicine practice practice. Prior investigations have selected a high risk population to compare XR to arthroscopic evaluation of arthritis. Therefore, we initiated a prospective comparison of standing AP and Rosenberg PA films in all patients presenting to a sports medicine practice who had subsequent arthroscopic surgery.

Methods:
As part of a prospective protocol all new patients had bilateral AP and Rosenberg PA x-rays, 349 patients subsequently had arthroscopic grading of arthritis as chondromalacia (CM) grade 1-4 (Modified Outerbridge) on scale diagrams by a single surgeon. All x-rays had articular cartilage interval independently measured in millimeters (mm) by compartments in both knees. Group A comparison was identical to Rosenberg with a positive radiograph as _2mm difference side to side (by compartment) for CM 0, 1, 2 vs 3, 4. In Group B only the definition of disease was changed with grade 2 CM now grouped with CM 3 and 4. The statistical evaluation compared sensitivity and specificity between x-ray procedure with p<0.05 as significant.

Results:
For Rosenberg method of analysis (see Table) in Group A only the specificity of PA view in the lateral compartment was significantly better (96% vs 92%). In Group B (see Table) the sensitivity of AP (11%) is significantly greater than PA (2%) for lateral compartment. In both analyses these findings are probably not of clinical significance given 4% difference in specificity or 9% change with extremely low values for sensitivity.

Conclusions:
Both radiography techniques (AP vs PA) demonstrated extremely low sensitivity but high specificity for arthritis whether analyzed by Rosenberg definition CM 1-2 vs CM 3-4 (Group A) or CM 1 vs CM 2, 3, 4 (Group B). Thus radiographs are poor screening tests for arthritis (2-4) (poor sensitivity) but are good diagnostic tests for chondromalacia with >90% specificity. No clinical difference or consistent statistical difference was observed between standing PA vs AP, therefore, either could be ordered with similar accuracy. Clearly a more sensitive screening tool is required for detection of early knee arthritis detected arthroscopically.