Although Grammont et al.²⁶ developed the reversed shoulder prosthesis nearly 20 years ago, it took two decades for this concept to be fully accepted in both Europe and North America. The reticence of many surgeons to accept this seemingly radical concept was because of the nonanatomic design and a long list of prior constrained prosthetic designs that led to dismal clinical failures. Currently, in Europe, the reversed shoulder prosthesis has become an important standard of care in the treatment of difficult reconstructive problems of the shoulder. Indeed, it has become one of the most commonly used prosthesis to treat rotator cuff tear arthroplasty and failed prior arthroplasty in the setting of soft-tissue loss. As such, it has greatly changed the habits of most surgeons who must confront these difficult problems. The reason for this great shift in clinical approach is the dramatic improvement of functional outcome because of the geometry of this prosthesis. In fact, the concept of “limited goal surgery,” by Neer et al.,³⁸ in the case of rotator cuff tear arthropathy and revision surgery has become antiquated in many cases because of the reliability of the reverse prosthesis in compensating for loss of rotator cuff function.

It is widely accepted that the outcome of an anatomic prosthesis depends on the status of the rotator cuff. Rotator cuff insufficiency often results in anterosuperior subluxation of the humeral head, which hinders deltoid action. The loss of a fixed fulcrum for rotation thus leads to loss of active elevation of the arm. Many surgeons attempted to design a constrained or reversed prosthesis that compensated for this biomechanical deficiency, but virtually all failed because of faulty biomechanical design. Grammont’s²⁶ legacy of the reversed shoulder prosthesis was based on his belief in the concept of reestablishing a fixed fulcrum for rotation while respecting the forces across the glenoid. His design allowed for both of these goals to be achieved. This chapter will discuss the history and biomechanical aspects of the reversed prosthesis as well as the surgical technique, indications, and results.

In the absence of an effective rotator cuff, active elevation is hindered because there is loss of the normal fulcrum for rotation that allows the deltoid to lift the arm. In fact, the orientation of the deltoid muscle fibers causes superior movement of the humeral head because the force coupled with the rotator cuff is now imbalanced. The coracoacromial arch forms the last barrier to this superior movement. The only way to restore shoulder motion in this situation is to restore a fixed fulcrum for rotation so the deltoid can have a sufficiently efficient lever arm to elevate the humerus. This can be achieved with the reversed shoulder prosthesis designed by Grammont.²⁶ Prior reversed shoulder designs have attempted to apply biomechanical principles normally appropriate for hip arthroplasty.⁸,⁹,³²,³⁴,⁴¹ These designs failed as the result of a high rate of glenoid loosening because they all moved the fulcrum for rotation lateral to its normal position and thus created increased shearing forces across the glenoid during elevation.⁴¹

The principle of a reversed prosthesis was suggested in the 1970s by Neer.³⁶ The advantage of a reversed design was to alter the vector of forces starting from the center of rotation to create centripetal forces, which would maintain the compression of the humeral cup onto the sphere during the elevation. The use of a large-diameter sphere was theorized to increase the mobility and lever arm of the deltoid by lateralizing the greater tuberosity.³,²⁰,³⁶ Neer observed,
however, that this type of implant would not reestablish active external rotation when the infraspinatus was torn. Thus, he abandoned this line of implant design in 1974. Moreover, all prostheses using a fixed rotation center outside the scapula were given up for the same reasons (Reeves prosthesis,⁴² Kolbel,³¹ Kessel,⁶,⁵² and Jefferson and Fenlin prostheses,²⁰ Gerard and Lannelongue prosthesis,²⁴ and Liverpool and Beddow prosthesis²). Three component prostheses suggested by Buechel et al.⁷ and Gristina and Webb²⁷ remained at the experimental stage and never progressed further.

In 1987, Grammont et al.²⁶ (Dijon, France) developed an original reverse prosthesis design, based on experimental biomechanical studies. The aim of Grammont et al. was to design a stable prosthesis by improving deltoid efficiency while limiting risk of glenoid loosening. According to the work of Fischer and coworkers,²¹ a prosthesis was conceived in which the center of rotation remained inside the glenoid. This geometry was achieved by fixing a hemisphere to a baseplate covering the glenoid to allow lateralization of the deltoid lever arm while at the same time reducing shearing forces across the glenoid and promoting compression forces across the joint during deltoid contraction and arm elevation (Figs. 26-1, 26-2 and 26-3).

The increase of the deltoid lever arm (torque) is caused by the medial displacement of the center of rotation, and, thus, an increase in the force magnitude exerted by the muscle during elevation (Torque = force × distance) (Fig. 26-2). A study on an anatomic simulator showed that a 10-mm medialization of the center of rotation increased the deltoid torque by 20% to 60%.²⁶ Because of the orientation of the force vector produced by the deltoid, lowering the center of rotation also induces an increase of the lever arm (lowering of 10 mm increases the abduction torque by 30%-60%). The increase of the deltoid forces is related to the middle deltoid tensioning, which improves the efficiency of the contraction of the muscular fibers. Furthermore, the medialization of the center of rotation also increases the abduction component of the anterior and posterior part of the deltoid (Fig. 26-3).

Using a large-diameter sphere allows an increase in the range of motion at the level of the prosthesis and increases the stability of a constrained design similar to the design of hip arthroplasty. The choice of a 36-mm or 42-mm sphere corresponds, according to the designer, to the average volume of the joint.¹ Beyond 45 degrees, the constraints applicable to the glenoid are centered on the bone/implant interface and participate in joint stability. The medialization and lowering of the rotation center tends to incline the diaphysis and open the angle between the humeral shaft and the scapula pilar. This tends to diminish the shearing constraints, which are persistent at the start of the elevation. For Grammont et al.²⁶ the existence of these constraints justifies the use of a noncemented glenoid implant with a central peg and a divergent fixation by screws fixed to the baseplate.

The initial prototype used by Grammont et al.²⁶ consisted of a cemented glenoidal implant corresponding to two-thirds of a sphere 42 mm in diameter, which would fit on a cylinder prepared with a crown saw. The humeral implant was polyethylene, forming a cup corresponding to one-third of a sphere (Fig. 26-4).

In 1987, Grammont et al.²⁶ reported their preliminary results on eight cases with an average follow-up of 8 years. There were three cases of postradiotherapy necrosis, four cases of revision prostheses, and one case of inflammatory arthritis. The approach included an osteotomy of the acromion. In three cases, patients recovered an active elevation of 100 to 130 degrees, and the author emphasized the swiftness of the functional recovery. From 1989 onward, Grammont conceived the Delta 3, in which the initial model consisted of a sphere screwed on the glenoid
baseplate (Fig. 26-4). Because of the complication of unscrewing the sphere, the fixation of the sphere was modified in 1995 by a Morse taper locked by a central screw. The humeral stem was conical and available in its cemented or uncemented form, covered by hydroxyapatite, and in three different lengths (100, 150, or 180 mm). The proximal part was screwed on the stem and designed with an antirotation fin. The humeral cup was made available in two thicknesses (0 and 6 mm) with the possibility of adding a metallic extension of 9 mm or using deeper versions in case of instability. The Delta prostheses 1 and 2 correspond to an anatomic version of the prosthesis with a metallic head to be fixed on the humeral implant (Delta 1) and an anatomic glenoid implant (Fig. 26-5).