Themistocles Gluck, a Romanian surgeon in the latter half of the 19th century, was the first to describe a prosthetic replacement as a potential treatment option in the shoulder.¹ His initial design was fabricated out of ivory and was anticipated to be used to treat tuberculosis infections in the shoulder. Unfortunately, there is no documentation that it was ever implanted in a patient.

The first recorded shoulder replacement was performed in 1893 by the French surgeon Jules Emile Péan for the treatment of tuberculosis in a 37-year-old baker at the Hôpital International in Paris.² The prosthesis design was originally inspired by Gluck, and it was fabricated by J. Porter Michaels, a dentist from Paris. The humeral stem was made of platinum that was connected by a wire to a rubber head coated with paraffin, creating a constrained implant (FIGURE 1.1). Although the patient initially had satisfactory results, the prosthesis eventually needed to be removed 2 years later because of a recurrent tuberculosis infection.

Nearly 60 years later, Frederick Krueger reported the use of a more anatomic shoulder prosthesis (FIGURES 1.2 and 1.3) constructed out of vitallium and molded from the proximal humeri of cadavers.³ This was successfully used to treat a young patient with osteonecrosis of the humeral head. Despite the success of this implant, it did not appear to gain any significant traction until a few years later. The design and development of the early modern-day shoulder replacements was pioneered by Dr. Charles Neer. In 1953, in response to his dissatisfaction with the outcomes of the operative management of proximal humeral fractures, Neer developed the first anatomic humeral prosthesis (FIGURE 1.4) for the treatment of a humeral head fracture.⁴ This was followed by the development of a second design referred to as the Neer I prosthesis. This first-generation implant was a monoblock design made of vitallium, with a fixed inclination and a straight cylindrical stem that attempted to replicate the proximal humerus anatomy. In 1955, he reported his results in 12 patients who were treated for acute four-part proximal humeral fractures, fracture-dislocations, and in patients with osteonecrosis from prior fractures.⁴ His results documented that 11 of 12 patients were pain free following this procedure. This led to an increased enthusiasm for the use of proximal humeral replacement to treat glenohumeral disorders. In 1964, Neer published a report describing the use of humeral head replacement for patients with a variety of glenohumeral disorders including acute fractures, fracture sequelae, and glenohumeral arthritis.⁵ This initiated a period of significant expansion in the development of shoulder replacement designs.

The early designs by Neer and Kruger represented nonconstrained implants that recognized the underlying biomechanics of the glenohumeral joint. The humeral head replacement was eventually joined with a glenoid component to begin the era of total shoulder arthroplasty. At the same time, other implants were being developed by shoulder specialists around the world that represented more of a constrained design. For some period of time, constrained designs were being developed for use in all types of glenohumeral arthritis and not necessarily only for patients with underlying rotator cuff deficiency. Our discussion of the development of shoulder arthroplasty will follow two tracks: the development of unconstrained designs and the development of fixed-fulcrum or constrained designs.

In the early 1970s, the design and implantation of glenoid components initiated the era of total shoulder arthroplasty. Neer designed a new humeral head replacement. The Neer II was characterized by a longer stem to dissipate the forces throughout the proximal humerus.⁶
It consisted of a spherical head with a 22-mm radius that was available in two head heights of 15 and 22 mm (FIGURE 1.5). The head had smooth rounded contours to accommodate the overlying rotator cuff tendons. This “monoblock” component is generally referred to as the “first-generation” shoulder arthroplasty. Neer also designed an all-polyethylene keeled cemented component that was rectangular in shape with a radius of curvature that matched the humeral component. In the 1970s and 1980s, the Neer II total shoulder arthroplasty was probably the nonconstrained shoulder arthroplasty system most commonly used for the treatment of traumatic, posttraumatic, and degenerative disorders of the glenohumeral joint.

The success of the Neer design led others to develop shoulder implant systems. The McNabb-English design (FIGURE 1.6) added more constraint to the articulation. The DANA design provided a nonconstrained option with a glenoid component that had a hood for increased constraint (FIGURE 1.7).⁷ Based upon the experiences with bipolar hip replacement, a bipolar shoulder design was also developed by Bateman. Nonconstrained implants had higher failure rates in the rotator cuff-deficient shoulder. This led Neer and others to develop more constrained designs, which will be discussed in the section on fixed-fulcrum designs.

Following the successful use of unconstrained designs, other implants were developed. The monoblock humeral component became a limiting factor because it did not address variations in humeral head size and dimension among different patients. The availability of one head radius (22 mm) and two head heights (15 and 22 mm) did not allow individual patient anatomy to be addressed. This occurred about the same time that modular components were being utilized for hip and knee replacement. Modularity then became part of total shoulder arthroplasty, which provided variable humeral head diameters and sizes that could be used to more closely match patient’s individual anatomy and enhance the stability of the construct.⁸ These modular components are generally referred to as “second-generation” implants (FIGURE 1.8). These designs allowed for customization in sizing of the humeral head and the ability to combine humeral stems of different sizes with humeral heads of different
sizes.⁸,⁹,¹⁰ Modularity improved the surgeon’s ability to match certain aspects of the patient’s individual anatomy, but it was soon recognized that it was not sufficient because the components did not account for the medial and posterior offset of the humeral head in relation to the shaft that characterizes proximal humeral anatomy.

In the 1980s, the anatomic studies by Boileau and Walch led to the development of “third-generation” implants.¹⁰ Based upon their anatomic studies, they described the proximal humerus as a sphere and a cylinder. A portion of the sphere represents the articular surface of the proximal humerus, and the humeral shaft represents a cylinder. They found that the diameter and thickness of the humeral head had a predictable relationship with each other but were highly variable between patients. In addition, they found that humeral neck inclination and retroversion were also highly variable. The humeral head (sphere) was offset posteriorly and medially from the humeral shaft (cylinder). These important findings led to the next phase of modularity,
with the design of eccentric humeral heads to provide a more anatomic restoration of the posterior and medial offset of the humeral head in relation to the shaft.⁸,⁹ These “third-generation” modular designs provided variable options for head thickness, offset, and diameter, which were all designed to achieve a more anatomic restoration for each specific patient (FIGURE 1.9). The principles behind the “third-generation” implants are for the prosthesis to be able to replicate the individual patient’s anatomy rather than using an implant with limited options, which requires the patient’s anatomy to adapt to the implant. his was a major step forward in implant design and resulted in improved patient outcomes.

Progression through second- and third-generation implants included a wide range of implant designs, including the Bateman bipolar prosthesis (FIGURE 1.10), Global Advantage (DePuy, FIGURE 1.11), Cofield, which included a noncemented glenoid (Smith and Nephew, FIGURE 1.12), Solar (Stryker), Bigliani/Flatow (Zimmer), and Aequalis (Tornier, FIGURE 1.13).

Further design changes led to the “fourth generation” of nonconstrained total shoulder arthroplasty characterized by increased modularity that allowed for in situ adjustability of the humeral head offset as well
as variation of inclination and version (FIGURE 1.14). Progression of shoulder arthroplasty designs from the first generation to the fourth generation has been based upon the goal of reproducing individual patient anatomy, which has clearly been shown to correlate with improved functional outcomes.

The successful use of press-fit standard length stems that relied primarily on metaphyseal fixation and some concerns about stress shielding and proximal humeral bone loss with longer stems led to the development of short humeral stems (FIGURE 1.15).¹¹ These short-stem designs maintained metaphyseal fixation and also offered the same modularity and anatomic restoration as standard stems. Their use has become widespread and, for the most part, successful. Further shortening of the stem led to the development of stemless implants (FIGURE 1.16). These were first introduced in Europe in 2004 and ultimately approved for use in the United States in 2015.¹² The benefits of short-stem and stemless implants include
bone preservation and the ability to resurface the proximal humerus when a proximal humeral deformity is present.¹³ Early results have certainly been encouraging, but longer term studies are needed to determine implant survival.¹²,¹⁴,¹⁵