BACKGROUND Two recent experimental studies evaluated the influence of shoulder rotation on neck-shaft angle, with conflicting results. However, there have been no clinical studies of whether this angle varies in different shoulder positions. The present study aimed to determine whether shoulder rotation affects neck-shaft angle on standard radiographs in patients with complaints of shoulder pain. HYPOTHESIS Shoulder rotation does not affect neck-shaft angle. MATERIALS AND METHODS A prospective study was conducted in selected patients with shoulder pain. Three true anterior-posterior radiographic views were obtained: in neutral rotation, 30° external rotation, and internal rotation with patient's arm in a sling. The X-rays were evaluated by three shoulder and elbow surgeons. Inter- and intra-observer reliability was evaluated by intraclass correlation coefficient (ICC). RESULTS Neck-shaft angle on true AP view did not differ between neutral rotation and 30° external rotation: 132±6° and 130±9°, respectively (P>0.999). In internal rotation with the hand resting on the abdomen, neck-shaft angle was 145±6°: i.e., significantly different (P<0.001) to the other two positions. Intra- and inter-observer correlation demonstrated excellent reliability. CONCLUSIONS Radiographic neck-shaft angle was significantly different in internal rotation with the patient's arm in a sling, compared with views in neutral or 30° external rotation. Intra- and inter-observer correlation showed excellent reliability. LEVEL OF EVIDENCE II, comparative prospective study.