Intraoperative fractures of the humeral shaft are closely associated with failure of the surgical technique such as inadvertent reaming, harsh impaction or inadequate manipulation of the arm during glenoid exposure [45, 80]. Spiral fractures are often caused by substantial torsional forces generated during external rotation of the shoulder. Inappropriate placement of the prosthesis or the reamer may finally result in a cortical perforation, most likely if the initial reamer or trial stem is not eccentrically positioned in the superolateral aspect of the proximal humerus [45, 80]. The presence of soft-tissue contracture as well as the necessity to remove a well-fixed cemented prosthesis for revision arthroplasty present challenges and may also result in intraoperative fracture [67].

In general, surgery is the management of intraoperative fractures of the humeral shaft. Simple cerclage wiring has been advocated for fractures proximal to the tip of the implant (Type A). If the tip of a standard prosthesis does not span the fracture site at least two to three cortical diameters, the implantation of a long-stemmed prosthesis is recommended [5, 80]. Similarly, intraoperative type B fractures should be treated with a long-stem prosthesis that spans the fracture site by two to three cortical diameters with the option of cement augmentation. Care should be taken to avoid extrusion of cement into the fracture site as this would impede healing [67]. Cerclages might be an option to augment fixation, especially in case of inferior bone quality (Fig. 28.3). Most periprosthetic fractures require at least extension of the deltopectoral approach to provide sufficient exposure of the fracture site. Fractures distal to the implant tip (Type C) warrant a long stemmed prosthesis placed through a combined deltopectoral and anterolateral humeral approach [80], alternatively plating could be an option. The distance of the fracture to the fossa olecrani is crucial and might prevent from sufficient stabilization using a long stem prosthesis. In case of very distal humeral shaft fractures additional plate fixation and/or cerclages should be considered. In patients in whom a standard cemented stem prosthesis has already been placed before recognition or generation of an intraoperative fracture, removal of the stem would risk extension of the fracture or nerve injury. In these cases, plate-and-screw and/or cerclage wire fixation adjacent to the stem is a viable option as well [67].

Postoperative fractures of the proximal humerus provide the option of non surgical treatment by bracing if they are minimally or non-displaced and if the component is well fixed [12]. Though, time to consolidation might be protracted and counts for an average of 180 days [38]. Postoperative type C fractures can be seen (and treated) as routine humeral shaft fractures which implements the possibility of a conservative regime in case of satisfactory alignment.

Loosening of the humeral stem would generally dispose the patient to surgical treatment and revision by long-stem implants. Both, cemented and non-cemented stems have been used in type A and B fractures in small case series and have shown satisfactory union rates [38, 81]. In case of a substantial overlap between the fracture length and a well fixed humeral stem (especially in type A fractures), as well as in case of a displacement of more than 2 mm and angulation greater than 20° in any plane, those fractures will preferably be treated as if the humeral component was loose [67]. Revision to a long-stem prosthesis is advised to bypass the fracture by at least two cortical diameters and fixation by cerclages should be supplemented distally. If necessary, plates and screws may afford torsional rigidity [67]. A displaced or unstable type B fracture with a well-fixed humeral stem is preferentially managed by plating. Recently, the locking compression plate (LCP) has been used with favorable outcome offering the possibility to combine wire-cerclages with plating, at least eight cortices should secure stability distal to the fracture [34]. Patients suffering from osteoporotic bone might benefit from additional use of allograft strut constructs [67]. Open reduction and internal fixation is recommended for fractures distal to the stem tip (Type C) without signs of healing after at least 3 months showing a well fixed stem, whereas revision with a long stem should be done for similar fractures associated with a loose humeral component [38].

The success of fracture-arthroplasty relies closely and more frequently than in elective surgery in the integrity of the tuberosities. Their acute dislocation or nonunion cause about 30 % of all complications in fracture arthroplasty and are one of the leading issues for revision surgery [76].

Fractures or involvement of the tuberosities may be treated by additional transosseous repair using non-absorbable sutures to secure the rotator cuff attachments [67], correct positioning of the fragments is crucial to achieve good and stable long-term results. Fixation of the greater tuberosity more than 2 cm distal to the apical circumference of the humeral head leads to overstuffing and deficient, functional results [48], whereas a cranialisation might cause subacromial impingement with limited function mostly in terms of limited abduction.

Only a stable and tension-free repair of the tuberosities will facilitate reintegration. Several prosthetic designs have been developed especially for fracture arthroplasty providing a better integration of the bony fragments in terms of the tuberosities and allowing for functional treatment postoperatively achieved either by a broad proximal shaping of the prosthesis enabling large interaction with the tuberosities or by the slim, so called “open stem design” offering multiple opportunities for transosseous suturing and good interfragmentary contact favoring ingrowth of the implant [76].

In the current literature it is stated that in primary fracture arthroplasty around 37 % of re-fixed tuberosities will reintegrate with a dislocation smaller than 5 mm, 17 % show more than 5 mm dislocation and the vast majority (46 %) will result in mal-positioning, non-union or bone resorption [37]. So the rate of expected complications and the need for revision-surgery because of fracture-involvement or intraoperative fracture of the tuberosities is imminent. Reintegration of the major tuberosity fragment has significant impact on the functional outcome, since more than 90 % of all patients presenting with less than 5 mm dislocation and adequate bone healing are satisfied with the surgical treatment result [37]. Furthermore, it has been proven that the patient’s age, the type of the implant and the total number of surgical procedures have a significant impact on the faith of the tuberosities, whereas the type of fixation seems negligible [37].

Once a fracture arthroplasty has failed because of dislocation of a reintegrated greater tuberosity (more than 5 mm), a corrective osteotomy should be considered [10]. In case of non-union, resorption or a non-functional rotator-cuff, a conversion to reversed shoulder-arthroplasty is recommendable, though the results are known to be inferior to primary reversed arthroplasty [10].

Glenoid fractures occur almost only during surgery, mostly during the preparation of the bone-stock preceding the implantation of the glenoid-component. These fractures may compromise stability of the component, especially when involving the scapular neck and may lead to early, symptomatic loosening of the prosthesis. As these conditions are very unfavorable, the situation of a fractured glenoid might easier be avoided than managed. Glenoid resurfacing is not advocated when bone support is questionable. As a salvage step, the remaining intact glenoid can be sculpted with a hand burr or glenoid reamer to match the radius of curvature of the humeral head component [46, 80]. In case of an intraoperative fracture, bone-grafting combined with a revision (metal back) glenoid component with wedge reinforcement and screws may be employed to restore stability of the bone and the implant [46, 80].

Achieving secure long term fixation of the glenoid component is the primary goal in total shoulder arthroplasty. In this context, the low strength and the small volume of bone of the glenoid vault are critical factors to secure fixation and finally limit the long term “survival” of the implant [54]. To date, the most common mid and long-term complication of total shoulder replacement is glenoid loosening causing postoperative pain, limited function, and the potential need for revision surgery [69].

The reported prevalence of radiolucencies at the cement-bone interface of the glenoid component ranges from 0 to 100 % and increases with time. Ten years after surgery most authors observe radiolucency in terms of radiographic signs of loosening in at least 80 % of the cases standing for migration, tilting or shifting of the component in 34 % [33, 66, 83]. Though, only a small fraction with radiologic signs of loosened glenoid components needs revision surgery (7 % after 13 years [5], 9 % after 10 years [83]). Efforts have been made to improve long term stability, including the preservation of the subchondral plate, concentric reaming of the glenoid, selected biomaterials and advanced prosthetic design [44, 69, 77]. Cemented pegged components were most commonly used and supposed to provide the most predictable fixation. In the current literature, pegged designs showed advantages in terms of better implant-seating and less signs of radiolucencies, finally leading to lower implant-loosening rates [23]. Furthermore curved back glenoids turned out to be beneficial concerning malpositioning-related failure however leading to higher mid- and long-term failure [71]. The technique and mode of cementation has significant impact on implant stability whereas a uniform cement mantle of 1 mm and implementation of a so called pressurization-technique show the best results [36, 72]. Non-cemented glenoid components on the other hand rely upon mechanical interlock and biologic integration, typically by screw fixation or a combination of screws and press-fit pegs, to achieve sufficient initial fixation facilitating bone in-growth. Although non-cemented glenoids offer many theoretical advantages compared to cemented glenoids, they have been associated with higher complication rates due to increased ultra-highmolecular-weight polyethylene (UHMWPE) wear and to joint-overstuffing [6]. Boileau et al. identified two major causes of metal-back glenoid loosening as follows: primarily the mechanical failure arising from a lack of initial stability and secondly osteolysis caused by PE and metal wear. In accordance, the four primary failure modes of metal-back glenoids were summarized as: (1) insufficient polyethylene thickness (4 mm instead of 5 mm); (2) Excessive thickness of the component (7 mm) with massive stressing of the rotator cuff; (3) Rigidity of the metal-back component accelerating polyethylene wear and stress-shields bone; and (4) Posterior i.e. eccentric loads on the glenoid leading to polyethylene disassociation [6]. For these reasons, Boileau et al. concluded that the fixation of metal-back glenoids is inferior to that of cemented glenoids which is also confirmed by other authors in the literature [44, 68]. Resent trends follow a combination of the two strategies entitled as minimally cemented glenoids [69].

The common opinion about the mechanism of glenoid loosening is a repetitive, eccentric load of the humeral head onto the glenoid, commonly called “rocking horse” phenomenon. This eccentric or edge loading conditions produce a torque on the fixation surface inducing tensile stress at the bone-implant or bone-cement-implant interface, potentially causing interfacial failure and glenoid disassociation [69]. When the glenoid is resurfaced using conforming implants, eccentric loading is enhanced because of the inability of the artificial surface (PE) to mimic viscoelastic properties of the former cartilage and labrum as essentials of the physiologic gleno-humeral motion. Eccentric loading might also result from glenoid mal-positioning or seating and humeral mal-positioning, all finally leading to implant-failure. The radial mismatch has been introduced to cope this problem, showing the ideal compromise between stability and native kinematics with a mismatch of 6–7 mm [77]. Finally an intact rotator-cuff preserves the artificial joint, because the magnitude of eccentric loading increases with weakness or insufficiency of the rotator cuff.

If a loose glenoid-component has to be revised there are strong arguments pointing towards a re-resurfacing of the glenoid [3, 15, 20]. Three retrospective surveys by Antuna et al., Deutsch et al. and Cheung et al. compared the functional and subjective outcome of revision-procedures with and without glenoid reimplantation. All uniformly described a better symptomatic and functional outcome for the patient group with reimplanted glenoids [3, 15, 20]. It is important to note, however, that these findings may include a selection bias, since patients in whom a glenoid implant was possible had a better bone stock and soft tissue situation than those patients being revised with a hemiarthroplasty.

The glenoid bone stock is crucial when considering re-resurfacing after aseptic loosening. As in primary surgery, attention should be drawn to a correct orientation of the glenoid component which can be achieved by eccentric reaming. Though, the limits of correction are even smaller than in primary resurfacing. Gillespie et al. conducted a cadaveric analysis on eight specimens to evaluate the degree of glenoid retroversion that can be corrected with eccentric reaming in primary gleno-humeral arthritis. They reported that an anterior correction of 10° resulted in a significant decrease of the glenoid width, 15° anterior correction resulted in an inability to seat the glenoid in 50 % of the tested specimens due to an inadequate bone stock, and 20° anterior correction resulted in an inability to seat the glenoid in 75 % of the tested specimens [25]. These results led the authors to suppose that 10° of anterior correction may be the limit. Beyond these findings, bone grafting should also be considered. Accordingly, treatment strategies for glenoid revision should include bone grafting to improve the glenoid bone stock. In these cases reversed shoulder arthroplasty represents a reliable therapeutic option providing the benefit of glenoid bone stock reconstruction by fixing the bone graft with a baseplate and screws and of solving the immanent problem of soft tissue insufficiency and prosthetic instability [4, 47]. Patients, in whom intraoperative findings preclude immediate component reimplantation would be candidates for singular bone grafting. Repeated revision with a glenoid component after graft consolidation should be considered for patients with continuous pain during shoulder activity after component removal and grafting [3, 56].

Aseptic loosening of the humeral component, despite its rare occurrence, accounts however for approximately 7 % of all reported complications in refer to shoulder arthroplasty [5]. Similar to loosening of the glenoid component, radiolucent lines on radiographs are the first indicator of ongoing disintegration of the implant. Recent reports have indicated a higher frequency when press-fit humeral stems were used [45]. Thus, humeral component survival was supposed to be affected by the chosen type of fixation and the biologic response to wear particles. Changes at the periprosthetic humeral interface in the presence of a glenoid component raised concern about osteolysis and the potential for symptomatic loosening [59]. Sperling et al. defined humeral components at risk if they showed radiographic evidence of subsidence, tilting or lucent lines of more than 2 mm around the implant [66].

Male gender, younger age (>65 years) and acute fracture or posttraumatic arthritis have been associated with an increased risk of revision surgery due to loosening of the humeral component [18]. Besides, the survival of the humeral component (cemented or press-fit) seems significantly decreased in prosthetic settings using metal-back glenoid components [18]. These findings are in accordance with biomechanical studies reporting high stress within the polyethylene of metal-backed glenoid components, with the implication that these components have inferior wear properties [70].

If a humeral component is loose it might be revised by either standard or long-stem implants. In cases with poor bone-quality the revision with a long-stem component is recommended. Bone insufficiency might additionally require the use of allografts or extension towards tumor-prosthetic reconstructions, both procedures highly associated with a poor clinical and functional outcome [2, 13].