Reliability of Visual, Goniometer, and Radiographic Assessment for Thumb Metacarpophalangeal Joint Hyperextension: An Analysis of Inter and Intraobserver Reliability
Nathaniel B Hinckley, DO, Sailesh V Tummala, MD, Kevin J Renfree, MD, Molly Klanderman, PhD and Heidi Kosiorek, MS, Mayo Clinic Arizona, Phoenix, AZ
Background: Symptomatic basilar thumb arthritis is often associated with "collapse deformity"-flexion of the thumb metacarpal and hyperextension of the metacarpophalangeal joint (MCPJ). When performing a basal joint arthroplasty, many authors have advocated for a capsulodesis or arthrodesis of the thumb MCPJ if >30 degrees passive hyperextension, in order to improve pinch biomechanics. There may be considerable variability in measurement methods and interpretation.
Purpose: To analyze the inter- and intraobserver agreement of thumb MCPJ hyperextension measurements utilizing visual estimates, goniometry, and radiographic assessment. Our null hypothesis was that visual and goniometric measurements vary greater than 10 degrees from the "gold standard" radiographic measurements (intra-observer agreement) as well as between observers (inter-observer agreement).
Methods: Twenty-six hands from 13 fresh-frozen cadaveric specimens were measured by an orthopaedic surgery resident (Rater A) and fellowship trained hand surgeon (Rater B). Each rater assessed maximal passive thumb MCP joint hyperextension by: 1. visual estimation, 2. goniometric measurement, and 3. axis measurement using a radiographic true lateral thumb image. Prior to the study, raters agreed upon technique of performing each measurement, but were blinded to each other's and their own prior measurements. Descriptive statistics were recorded for measurement type as well as the interobserver reliability using a two-way intra-class correlation (ICC). Intra-observer reliability was calculated using the concordance correlation coefficient (CCC) for each observer. Bland-Altman plots were utilized to identify trends, systemic differences, or potential outlier for both analyses.
Results: The mean measurement for the two raters was similar for visual estimation and radiographic measurement. Using goniometry, the average measurement was 2x higher for Rater B than A (18.0 vs. 9.8). Across both raters, radiographic measurements were 10 degrees greater than the other two methods. For inter-rater agreement, measurements were within 10 degrees: visual -92%, radiographic -96%, goniometer- 68%. Rater B tended to have better agreement between visual and goniometric measurements compared to radiographic measurements.
Conclusions: Radiographic assessment portends a greater interobserver agreement for thumb MCP joint hyperextension than visual estimate or goniometric measurement. Rater experience may improve precision, but there is still poor agreement between visual estimates and goniometer measurements when compared to radiographic measurements, as the former two tend to underestimate the actual amount of hyperextension by 10 degrees.
Back to 2022 ePosters