Shoulder replacement arthroplasty.

Since 1973 64 shoulders in 56 patients have had a prosthetic replacement at the Oslo Sanitetsforening Rheumatism Hospital, the semi-constrained prosthesis of Lettin/Stanmore was used in 13 cases, the non-constrained prosthesis of Engelbrecht/St. Georg in 10 cases and Neer's type II prosthesis in 41 cases. An evaluation of the results is given. The minimal muscular loosening recommended by Neer is facilitating the postoperative training. The best results regarding function and pain reduction have been obtained by the Neer prosthesis, but the lack of cranial support ("fornix humeri") and the ovalization of the head seem to be responsible for some cases of upward migration of the head of the prosthesis. The ideal prosthesis should be non-constrained and come with different lengths of the neck and at least in two sizes of the head and several sizes of the stem. The glenoid socket should provide a cranial support facilitating the centering. Special problems are met with in juvenile rheumatoid arthritis. A best possible reconstruction of the rotator cuff and the deltoid is mandatory, requiring good technical skill.