The objective of this study was to determine the effect of different prosthetic systems on the functional and radiographic outcomes after shoulder arthroplasty for fractures. This study comprised 35 patients (28 women and 7 men) with a mean age of 74 years (range, 56-88 years) who sustained 4-part fractures of the proximal humerus and were randomly allocated to 2 different groups regarding the type of prosthesis. The 2 systems used differ mainly in the type of fixation of the tuberosities. In group 1 (EPOCA), the fixation was achieved with wire cables through a medial and a lateral hole in the stem, whereas in group 2 (HAS), the fixation was performed by use of transosseous braided sutures. After a follow-up of 1 year, the functional and radiographic outcomes were evaluated. The retrieved data demonstrate that rigid fixation and anatomic positioning of the tuberosities (group 1) increase the rate of bony healing superior to all other factors. There was a statistically significant difference regarding the relative individual Constant score (P = .001) and the mean active range of motion (flexion, P < .001; abduction, P = .001; external rotation in adduction, P = .01; and external rotation in 90 degrees abduction, P = .001) when both groups were compared, showing a better outcome in the EPOCA group for all parameters. Radiologic findings, like heterotopic ossification, glenoid erosion, or subluxation, had no significant influence on the outcome in this study. Accurate placement of the tuberosities and healing at the bone-bone interface of the rotator cuff seem to be the most important factors influencing the outcome in prosthetic care of fractures.