Multirater agreement of arthroscopic grading of knee articular cartilage

Robert G. Marx, Jason Connor, Stephen Lyman, Annunziato Amendola, Jack T. Andrish, Christopher Kaeding, Eric C. McCarty, Richard D. Parker, Rick W. Wright, Kurt P. Spindler

Research output: Contribution to journalArticlepeer-review

91 Citations (Scopus)


Background: Acute and chronic cartilage injury of the knee has an important impact on prognosis. The validity of the classification of such injuries is critical for prospective multicenter studies. The agreement among multiple surgeons at different institutions for articular cartilage lesions has not been established. Hypothesis: Arthroscopic classification of articular cartilage lesions is reliable and reproducible and can be used for multicenter studies involving multiple surgeons. Study Design: Cohort study (diagnosis); Level of evidence, 1. Methods: A total of 6 surgeons from 5 centers reviewed 31 videos of articular cartilage lesions. With grade 2 and grade 3 combined for the analysis, observed agreement ranged from 81% to 94%, and kappa ranged from 0.34 to 0.87. An additional 22 videos comprising grade 2 and grade 3 lesions were analyzed, and the observed agreement was 80%, with an overall kappa of 0.47. Conclusion: Arthroscopic grading of articular cartilage lesions is reproducible among surgeons at different centers. Clinical Relevance: Articular cartilage lesions can be reliably classified among surgeons at different sites. Such reliability is important for multicenter clinical research studies involving arthroscopic knee surgery.

Original languageEnglish
Pages (from-to)1654-1657
Number of pages4
JournalAmerican Journal of Sports Medicine
Issue number11
Publication statusPublished - Nov 2005
Externally publishedYes

All Science Journal Classification (ASJC) codes

  • Physical Therapy, Sports Therapy and Rehabilitation
  • Orthopedics and Sports Medicine


Dive into the research topics of 'Multirater agreement of arthroscopic grading of knee articular cartilage'. Together they form a unique fingerprint.

Cite this