Dorsal/volar tilt is an important radiographic measurement commonly included in the treatment decision of distal radius fractures (DRFs). However, studies have shown that forearm positioning relative to rotation (i.e., supination and pronation) can affect the measured tilt value but with substantial interobserver variance. To examine whether interobserver agreement on radiographic tilt measurement is influenced by forearm rotation. We radiographed 21 cadaveric forearms at 5° rotational intervals between 15° supination and 15° pronation on lateral radiographs. A radiologist and a hand surgeon measured tilt in a blinded and randomized fashion. Bland-Altman analyses with bias and limits of agreement (LoA) were used to estimate interobserver agreement for forearms in all degrees of rotation, non-rotated forearms, supinated forearms, and pronated forearms. Interobserver agreement varied with forearm rotation. Bias (95% confidence interval [CI]; LoA) was -1.54° (95% CI: -2.53, -0.55; LoA: -13.46, 10.38) when measuring tilt on radiographs with all degrees of forearm rotation, and -1.48° (95% CI: -4.13, 1.17; LoA: -12.88, 9.92) when measuring tilt on true lateral 0° radiographs. When measuring on supinated and pronated radiographs, bias was -0.03° (95% CI: -1.35, 1.29; LoA: -8.34, 8.28) and -3.23° (95% CI: -5.41, -1.06; LoA: -16.90, 10.44), respectively. Interobserver agreement on tilt was similar when comparing measurements made on true lateral radiographs to those made on the group with all degrees of forearm rotation. However, interobserver agreement improved with supination and worsened with pronation.
Read full abstract