Authors:
T. Luijkx1, M. Maas2; 1Amsterdam, Utrecht/NL, 2Amsterdam/NL
DOI:
10.1594/essr2015/P-0116
Methods and Materials
72 wrist MRI scans of 55 consecutive patients were retrospectively evaluated for wrist position (defined by radio-ulnar angle,
RUA),
degree of ECU subluxation and the presence of ulnar-sided symptoms and MRI abnormalities.
Variations among one institution's MRI scanners were also evaluated separately.
A linear regression model was constructed quantifying the effect of wrist positioning on subluxation,
corrected for MRI abnormalities and symptoms,
age and gender.
Measurement of wrist position
The radio-ulnar angle was measured between a line along the ulnar groove and a line drawn between the centres of radius and ulna,
resulting in a quantitative measure of supination/pronation (Figure 1).
A cut-off of -15º (supination) and +15º (pronation) was used to create three discrete groups of pronation,
supination and neutral wrist positioning for comparison purposes.
Measurement of ECU position
The ulnar groove width was measured between the most protruding osseous edges of the ulnar groove.
The most medial (ulnar) edge of the ECU was used as the reference point for ECU position.
Displacement of the ECU was determined by the distance between this point and the medial (ulnar) edge of the ulnar groove,
both in absolute measures and as a percentage of the ulnar groove width (Figures 2 and 3).
Statistical analysis
Data on ECU displacement for three wrist positions (supination,
neutral,
pronation) were compared for differences in mean ECU displacement and absolute displacement as defined by ECU displacement larger than zero.
A Pearson correlation between percentage ECU displacement and radio-ulnar angle was calculated.
A linear regression model was constructed to explain variation in the percentage of ECU displacement in relation to the radio-ulnar angle (and therefore wrist position),
taking into account as covariates age,
gender,
presence of ulnar-sided symptoms and presence of ECU abnormalities on MRI. Variables were entered and consequently removed by backward selection one at a time when exceeding a p-value of 0.15.