Are acromiohumeral distance measurements on conventional radiographs reliable? A prospective study of inter-method agreement with ultrasonography, and assessment of observer variability.

Objectives: This study aims to investigate the reliability of acromiohumeral distance (AHD) measurements using conventional radiographs and to compare non-standardized and standardized radiographs with intra-/interobserver reliability measurements.

Patients and methods: Between February 2021 and January 2022, a total of 110 shoulders of 55 patients (25 males, 30 females; mean age: 49.7±12.6 years; range, 25 to 77 years) were included. Radiographs were taken in four different positions: primarily shoulder anteroposterior (AP), true AP, standardized true AP, and standardized outlet views. The AHD was measured by three orthopedists. A prospective ultrasonography (US) evaluation was performed by an experienced physiatrist, and the relationship between US and radiographic measurements was evaluated. The intra- and interobserver reliability of radiographic measurements was assessed.

Results: On the standardized true AP view measurements, all observers showed a moderate to good agreement with US measurements (intraclass correlation coefficients [ICC]: 0.68-0.75). There was no significant difference between the AHD measurements of the senior orthopedist on standardized true AP and outlet views, and the US measurements. The intraobserver agreement of US measurements was excellent (ICC: 0.98, 95% confidence interval [CI]: 0.98-0.99), and the intraobserver agreement level of measurements on radiographs were good to excellent with a wide range of ICC values (ICC: 0.79-0.97). Interobserver reliability was the highest on the standardized outlet view, with an ICC of 0.91 and 0.88 in two measurement times. Interobserver reliability of other measurements were good with ICC values ranging from 0.82 to 0.88.

Conclusion: The AHD measurements on radiographs are compatible with US measurements within up to 2 mm difference if standardization is ensured. Also, measurements on standardized views have a superior consistency with lower standard error of measurement and minimal detectable change values. Therefore, we recommend using standardized true shoulder AP and standardized outlet radiographs in clinical practice and studies, as these are the most accurate in demonstrating true AHD.