Objective: To evaluate the reproducibility and validity of the histologic/histochemical grading system (HHGS) for osteoarthritis (OA) developed by Mankin and coworkers.
Methods: Sections of human articular cartilage from macroscopically normal and osteoarthritic femoral heads were graded by the HHGS (0-14 scale) twice by 3 observers, and intra- and interobserver reproducibility was determined.
Results: The exact intra- and interobserver reproducibilities ranged from 22% to 33% and from 13% to 25%, respectively. Intra- and interobserver variation ranged from -7 to 5. Although scores were generally higher in cartilage from osteoarthritic joints as compared with cartilage from normal joints, the difference reached statistical significance for only 2 observers. Additionally, the average scores from all observers were within the definitive range for OA (>5) in 30% of sections from normal joints, and in only 47% of sections from OA joints.
Conclusion: The results indicate that the reproducibility and the validity of the HHGS are inadequate, and an improved histopathologic grading system for OA is needed.