The purpose of this study was to identify possible causes for the low union
rate for surgically stabilized os acromiale. Between February 1990 and Nov
ember 1995, fusion of an os acromiale was attempted in 15 shoulders in 12 p
atients at our institution. All patients were men. The mean age was 54 year
s (range 37 to 63 years). All shoulders had an os mesoacromiale considered
too large for simple resection. An associated lesion of the rotator cuff ra
nging from partial-to full-thickness tear was present in all patients. Elev
en had an unfused acromial epiphysis in both shoulders. Two different surgi
cal approaches were used. Seven shoulders were approached through an anteri
or deltoid-off approach, thus potentially devascularizing the os acromiale.
Eight shoulders were approached transacromially, preserving the deltoid or
igin and hence the terminal branches of the thoracoacromial artery. The tec
hnique of internal Fixation (tension band wiring) was the same for both gro
ups. The mean follow-up was 44 months (range 13 to 72 months). Union, as de
monstrated by axial radiographic views, occurred in 3 out OF 7 cases With a
devascularized os acromiale and in 7 Out Of 8 shoulders with a perfused os
acromiale (P = .017), respectively. Patients with a united os acromiale ha
d a significantly better functional outcome as measured by the Constant sco
re (P = .0169). In conclusion, aiming at a stable fusion of a sizable and h
ypermobile os acromiale is probably desirable because it enhances the overa
ll functional result. Obtaining consolidation was possible when the vascula
rity of the acromial epiphysis was respected.