Study design: A reliability study was conducted. OBJECTIVE.: To modify a grading system for lumbar disc degeneration and to test the reliability of this modified grading system.
Summary of background data: The 5-level Pfirrmann grading system for disc degeneration did not prove discriminatory when used to assess disc degeneration in the elderly spine. Such discriminatory power is necessary to test the association between other variables and severity of disc degeneration.
Methods: An 8-level modified grading system for lumbar disc degeneration was developed including a description of the changes expected for each grade and a 24-image reference panel. The reliability of the modified grading system was tested on 260 lumbar intervertebral discs in 52 subjects (26 men, 26 female) with a mean age of 73 years (range, 67-83 years). All examinations were analyzed independently by 3 readers. Intraobserver and interobserver reliabilities were assessed by calculating weighted kappa statistics.
Results: On average, for all 3 readers, 0.39% of the 260 discs were classified as Grade 2, 22% were classified as Grade 3, 21.5% were classified as Grade 4, 25.3% were classified as Grade 5, 19.1% were classified as Grade 6, 7.1% were classified as Grade 7, and 4.8% were classified as Grade 8. Intraobserver agreement was excellent (weighted kappa range, 0.79-0.91) with substantial interobserver agreement (weighted kappa range, 0.65-0.67). Complete intraobserver agreement was obtained, on average, in 85% of all discs with 84% of disagreement being as a result of a 1 grade difference. Complete interobserver agreement was obtained, on average, in 66% of all discs with 91% of disagreement being as a result of a 1 grade difference.
Conclusion: The modified Pfirrmann grading system is useful at discriminating severity of disc degeneration in elderly subjects. The system can be applied with good intra- and interobserver agreement.