• Rich JA
• Linaker KL
• Little JW
• Laptook EE

• Association of Chiropractic Colleges / congresses

OBJECTIVE: To evaluate the reliability of a 5-point scale for grading degenerative changes of the lumbar zygapophysial (Z) joints from standard radiographs.

STUDY DESIGN: A retrospective inter-rater and intra-rater reliability study involving 3 examiners.

METHODS: A modified protocol for Kellgren’s classification of degenerative changes of the Z joints was evaluated for reliability in the lumbar spine. Seventy-nine individual Z joints were independently graded by two primary examiners on two separate occasions separated by three weeks. A third examiner (a radiologist) was also used as an expert for inter-rater reliability comparisons between measurements made in sessions 1 and 2 for both primary examiners. Percent agreement and weighted Kappa scores (KW) were calculated to determine inter-rater and intra-rater reliability.

RESULTS: Inter-rater agreement between examiner 1 and 2 was KW =0.60 with 49.4% perfect agreement (39 levels) and 40.5% one grade difference (32 levels) for a total of 89.9% agreement within one grade. Intra-rater agreement was lower than inter-rater agreement. Examiner 1 (KW =0.42) demonstrated 39.2% perfect agreement (31 levels) and 40.5% agreement with one grade difference (32 levels), resulting in 79.7% agreement within one grade. Examiner 2 (KW =0.54) demonstrated 33.75% perfect agreement (27 levels) and 52.5% agreement with one grade of difference (42 levels) for a total of 86.25% agreement within one grade.

Inter-rater reliability was also evaluated and compared for both grading sessions of examiners 1 and 2 with examiner 3. In session 1, agreement was KW = 0.37, with 61.9% agreement within one grade (27.8% perfect agreement) for examiner 1 and examiner 3, and KW =0.39 with 69.5% agreement within one grade (16.4% perfect agreement) for examiner 2 and examiner 3. Session 2 revealed KW =0.57, with 94.9% agreement within one grade (perfect agreement 32.9%) for examiner 1 and examiner 3, and KW =0.68 with 94.9% agreement within one grade (50.6% perfect agreement) for examiner 2 and examiner 3. The average KW for session 2 was 0.63, which is a substantial strength of agreement.

CONCLUSIONS: The Modified Kellgren five-point grading system used in this study provides acceptable intra- and inter-rater reliability when observers are adequately trained. The Modified Kellgren grading system may be a useful method for future investigations assessing the relative degeneration of the Z joints in research subjects and may also be useful in assessing Z joint degeneration in back pain patients.

ACKNOWLEDGEMENTS: We would like to thank Dr. Ray McKinnis for his work in the statistical analysis of this paper. We would also like to thank the collaborative efforts of the NUHS Departments of Research and Diagnostic Imaging who made the completion of this project possible.

This abstract is reproduced with the permission of the publisher.