Observer variability in the assessment of the acromiohumeral interval using anteroposterior shoulder radiographs.

Abstract

An acromiohumeral interval narrower than 6 mm measured on AP shoulder radiographs has been considered pathological and suggestive of rotator cuff tears. This prospective study was conducted to assess inter- and intraobserver variation in the radiographic assessment of the acromiohumeral interval and its critical value on routinely taken AP shoulder radiographs off-study use to evaluate the accuracy of this measurement method. The acromiohumeral distance from the inferior anterior acromial aspect to the humeral head was measured in millimeters. Thirty blinded, anteroposterior shoulder radiographs were independently reviewed by five board-certified orthopedic shoulder surgeons at two time points in random order. The results of three investigators showed significant intraobserver variation. Five investigator pairs showed significant interobserver variation at both examination time points. The maximum interobserver difference for the same radiograph was 8 mm (range 0 to 8 mm). Our results indicate that the assessment of the acromiohumeral interval using non-standardized anteroposterior radiographs off-study use cannot be seen as a reproducible and reliable method of measurement.