RADIOGRAPHIC MEASUREMENTS OF THE ACROMIOHUMERAL INTERVAL ON STANDARDIZED ANTEROPOSTERIOR RADIOGRAPHS ARE NOT ACCURATE; A PROSPECTIVE STUDY OF OBSERVER VARIABILITY

Abstract

Background: An acromiohumeral interval narrower than six millimeters has been considered pathologic and strongly indicative for rotator cuff tears by numerous authors. This prospective study was conducted as an assessment of inter- and intraobserver variation in the radiographic measurement of the acromiohumeral interval and its critical value.

Material and Methods: Thirty blinded, standardized anteroposterior shoulder radiographs were independently reviewed by five board certified orthopedic shoulder surgeons. The acromiohumeral distance, from the inferior anterior acromial aspect to the humeral head was measured in millimeters. The five investigators classified each film a second time in random order.

Results: Finally the same 27 radiographs (90%) have been evaluated by five investigators at both examination time points. The results of three investigators showed significant intra-observer variation ranging from 4 to 11 millimeters for the same radiograph. Six investigator pairs showed significant inter-observer variation at both examination time points (p< 0.05). The maximum inter-observer difference for the same radiograph was 11 millimeters (ranging from 1 to 11, SD 0.3 – 4.2).

Conclusion: In view of our results the assessment of the acromiohumeral interval using anteroposterior x-rays does not seem to be a reproducible method of measurement. Further investigations in combination with CT or MRI are necessary to ensure our findings.