Prevention and management of reverse shoulder arthroplasty complications: A structured current concepts review

Abstract

Recently, reverse shoulder arthroplasty (RSA) has been implemented in a variety of shoulder disorders including revision surgeries, challenging proximal humeral fractures, post-traumatic sequelae, or the classic rotator cuff tear arthropathy. With the objective of formulating prevailing preventative and treatment strategies, this review offers a rigorous, evidence-based appraisal of RSA complications by incorporating recent clinical literature and long-term worldwide registry data. Problems are arranged chronologically into three separate phases to offer a useful clinical framework: (1) Intraoperative mechanical problems, with an emphasis on metaphyseal fractures and component mispositioning; (2) Early postoperative including acute instability, neurological damage, and early periprosthetic infection; and (3) Long-term biological and fixation failures represented in scapular notching, chronic aseptic loosening, and stress shielding. Importantly, this review highlights how these complication profiles in all three phases have been substantially altered through the widespread biomechanical shift from traditional medialized designs to contemporary lateralized configurations, which has mitigated late scapular notching while concurrently increasing vulnerabilities to early and mid-term acromial stress fractures. Furthermore, recent algorithms that compare single-stage vs two-stage revision procedures for low-virulence infections that emerge late are thoroughly assessed. In the contemporary era, minimizing failures necessitates precision surgical execution, an extensive awareness of implant biomechanics, and patient-specific risk factors. To optimize patient selection, mitigate intraoperative risks, and properly address postoperative challenges, this review aims to demonstrate systematic timeline-based algorithmic strategies.

Key Words: Reverse shoulder arthroplasty; Complications; Instability; Scapular notching; Periprosthetic infection; Lateralization; Revision arthroplasty; Acromial stress fracture

Core Tip: As clinical indications for reverse shoulder arthroplasty continue to expand globally, navigating associated failures requires a highly systematic approach. This structured narrative review provides a practical timeline-based framework for the prevention and management of reverse shoulder arthroplasty complications. Such approach categorizes complications into intraoperative, early, and late postoperative phases. This article details how modern implant design’s shifts have altered complication paradigms and deliver clear actionable clinical algorithms to guide surgeons through revision procedures.

INTRODUCTION

Reverse shoulder arthroplasty (RSA), developed by Paul Grammont, has progressed from being a salvage option for irreparable rotator cuff (RC) deficient shoulders to an exceedingly common type of shoulder replacement performed worldwide[1,2]. Indications have progressively broadened far beyond cuff tear arthropathy (CTA). Indications include acute complex proximal humeral fractures (PHFs), post-traumatic sequelae, failed prior arthroplasty or RC repair, inflammatory and primary glenohumeral arthritis with an intact RC, and oncological reconstruction after proximal humeral tumor resection[1,3,4]. National registries affirm this pattern with RSA accounting for more than 60%-80% of all shoulder arthroplasties performed each year in the United States and Australia[1]. Similarly, the National Joint Registry for England, Wales, and Northern Ireland collectively report an annual increase in case volume[5].

This expansion has been paralleled by an equally instructive accumulation of complication data. Earlier series reported an overall complication rate of 15% to 24%[6,7], a figure that, while still frequently cited, obscures considerable study-to-study heterogeneity driven by inconsistent complication definitions, variable follow-up duration, and the conflation of primary and revision indications within pooled estimates[6,8]. Contemporary registry-level data suggest that revision rates for primary RSA are, in absolute terms, comparatively low (in the range of 2%-5% at five to seven years)[9-11].

A further, and arguably the most consequential, development of the past decade has been the deliberate biomechanical migration of implant design away from the original Grammont medialized, semi-constrained configuration toward lateralized designs (glenoid and/or humeral components)[12-14]. This shift was undertaken specifically to reduce the scapular notching and loss of external rotation that characterized medialized designs[15-17], but it has, in turn, generated a new complication signature dominated by acromial and scapular spine stress fracture, altered instability patterns, and distinct fixation challenges at the humeral and glenoid interfaces[18-20]. Any contemporary review of RSA complications must therefore be organized not only by anatomical side (humeral vs glenoid) as in earlier literature[21,22], but by the interaction between the temporal course of failure and the biomechanical era of the implant used.

This review adopts a chronological three-phase organizing framework including intraoperative, early postoperative, and late biological/fixation-related complications. This framework mirrors the sequence in which the orthopedic surgeon encounters these problems in practice. It also integrates how the modern lateralized RSA designs have reshaped the incidence and character of each complication category. This framework is purposefully different from previous influential reviews, which have arranged RSA complications primarily by anatomical side (humeral vs glenoid) or by type of complication (mechanical vs biological) without a clear temporal axis[21-25].

Modern lateralized implants have shifted the complication timeline rather than just changing the type of failure. While medialized designs caused early scapular notching, lateralized options introduce delayed acromial stress fractures. Traditional anatomical or type-based classifications fail to capture this chronological shift for the operating surgeon. This chronological framework aligns directly with clinical decision-making phases. It guides surgeons from intraoperative technique through early postoperative surveillance for instability and periprosthetic joint infection (PJI). Finally, it extends to long-term radiographic monitoring for implant survivorship. This structure translates design-specific complication profiles into a practical, phase-based management protocol.

SEARCH STRATEGY

A complete literature search was conducted using PubMed, Scopus, and Web of Science databases. The search included articles till June 2026. The search strategy used combinations of the terms “Reverse Shoulder Arthroplasty”; “Reverse Total Shoulder Replacement”; “Complications”; “Instability”; “Dislocation”; “Scapular Notching”; “Periprosthetic Joint Infection”; “Periprosthetic Fracture”; “Acromial Stress Fracture”; “Aseptic Loosening”; “Lateralization”; “Revision Arthroplasty”.

Clinical research, systematic reviews, meta-analyses, and biomechanical investigations were prioritized. Orthopedic literature from the last five years was emphasized as it reflected the modern lateralized RSA. Furthermore, past studies were maintained in cases where they were the major source for a classification system. This is a narrative review of contemporary concepts, not a systematic one. Consequently, no formal PRISMA protocol was registered.

To enable reproducibility, the search and selection process mentioned above is transparent. Only full-text English-language publications that reported clinical, biomechanical, or registry-level data on RSA complications were eligible. Conversely, conference abstracts, case studies lacking broadly applicable conclusions, and non-English texts without a translation were not included. Systematic reviews, meta-analyses, registry-based studies, and studies published within the last five years were given priority for inclusion when multiple studies addressed the same question, unless an earlier investigation represented the original description of a still-in-use classification system.

EVOLUTION FROM MEDIALIZED TO LATERALIZED DESIGN

Grammont’s original design medialized both the glenoid center of rotation (COR) and the humeral socket (Figure 1), which lengthens the deltoid lever arm and lowers the joint reaction force but at the cost of bony impingement between the humeral polyethylene cup and the scapular neck during adduction (the substrate for scapular notching)[15,17,26]. Successive design iterations have lateralized the COR either at the glenoid [bony increased-offset (BIO)- RSA, metal augments, or eccentric glenosphere] or at the humerus [onlay humeral trays, inlay stems with 135° rather than 155° neck-shaft angles (NSA)], or both (bipolar)[13,14,27]. Meta-analytic data show that lateralized designs reduce scapular notching by approximately 2.5-fold and heterotopic ossification by approximately 2-fold relative to medialized designs[12]. Mid-term series of lateralized implants have reported notching rates approaching zero where medialized cohorts from the same institution previously reported rates near 70%[28].

Figure 1 Center of rotation of native shoulder and Reverse shoulder arthroplasty. A: Orientation of subscapularis muscle fibers in the native shoulder crossing the native center of rotation (COR) (orange circle); B: Altered orientation of muscle fibers through the new medialized and distalized COR (yellow circle) of reverse shoulder arthroplasty.

Nonetheless, this benefit is counterbalanced by a biomechanical cost. Glenoid sided lateralization shifts the COR away from the glenoid face and mainly improves impingement free adduction and internal and external rotation, at the cost of greater shear load across the baseplate bone interface. Humeral sided lateralization instead increases deltoid and RC remnant length and tension, and this is the change most consistently linked to acromial and scapular spine stress fracture[19,20,29]. Combined lateralization may compound both effects and appear to carry the most favorable notching profile alongside the least favorable acromial fracture profile. For this reason, the glenoid and humeral contributions to lateralization deserve separate consideration rather than being treated as one design parameter, particularly in patients with osteoporosis or preexisting acromial thinning, where the humeral contribution warrants closer scrutiny.

PHASE I (INTRAOPERATIVE MECHANICAL COMPLICATIONS)

Intraoperative complications are often demonstrated as technique-dependent complications. Thus, they are mainly modifiable through meticulous preoperative planning, implant selection, and surgical execution. Pooled data from a systematic review of over 13500 RSAs procedures found an overall intraoperative fracture rate of 1.4%, rising markedly to 13.6% in the revision setting compared with 2.9% in primary RSA[30].

INTRAOPERATIVE HUMERAL FRACTURES

Humeral shaft or metaphyseal fracture typically occurs during humeral canal preparation (broaching, reaming) or component impaction, and less commonly during joint reduction and manipulation[31-33]. Reported risk factors include female sex, osteoporosis, prior humeral hardware or deformity, a tight or sclerotic canal, and the revision setting, where removal of a well-fixed stem and cement mantle disproportionately drives the fracture rate[30,34]. In the aforementioned systematic review, 136 of 188 intraoperative fractures were humeral, 60% occurring in revision RSA during implant extraction, most commonly involving the diaphysis[30]. The classification proposed by Wright and Cofield[32] and later refined by Campbell et al[35] and by Kirchhoff et al[36] for the periprosthetic setting, stratifies fractures by location relative to the stem tip, and by stability. Figure 2 illustrates the three categories defined by Wright and Cofield[32] according to their anatomical configuration. Type A fracture can be managed conservatively. However, operative treatment should be considered for type B fractures. On contrary, Kirchhoff et al’s classification[36] (Table 1) relies upon both prosthesis stability, fracture localization, and height.

Figure 2 Illustration image for periprosthetic humeral fracture as per the Wright and Cofield classification. A: Located at the tip of the prosthesis and extends proximally; B: Mimics A without proximal extension; C: Distal to the prosthetic tip.

Table 1 Kirchhoff et al[36] classification for periprosthetic humeral fractures after reverse shoulder arthroplasty.

Humeral fracture	Prosthesis stability	Management option
1: Tuberosities	S: Stable	1-S: Conservative or ORIF
2: Spiral	L: Loose	2-S: ORIF
3: Oblique		3-S: ORIF
4: Distal		L: Revision by long stem RSA ± ORIF

ORIF: Open reduction and internal fixation; RSA: Reverse shoulder arthroplasty.

Intraoperative humeral periprosthetic fractures are better avoided than managed. Release of inferior capsular attachment to the humerus, avoiding excessive torque while positioning the arm in adduction and extension can guard against fracture development. Additionally, the humeral canal is better reamed by hand, broaching in line with the humeral shaft are suggested. However, excessive fitting of the humeral stem is better avoided. Furthermore, glenoid implantation should not be postponed after complete humeral preparation; as the spherical metaphyseal reamer usually leaves a thin humeral cortical shell, which is prone to fracture during glenoid exposure[23].

Stable nondisplaced fractures about a well-fixed stem can generally be managed with a period of protected weight-bearing and cerclage cabling alone, whereas displaced fractures, those associated with stem loosening, or fractures extending distal to the stem tip typically require open reduction with a plate construct spanning the fracture supplemented by cerclage wires, or conversion to a long-stem (revision) humeral component that bypasses the fracture by at least two cortical diameters[34,35,37,38]. Prophylactic cerclage cabling prior to broaching in osteoporotic bone, meticulous sequential reaming rather than forceful impaction, and preoperative templating of canal fit are the principal preventive measures[30,37].

INTRAOPERATIVE GLENOID FRACTURES AND BONE DEFICIENCY

Glenoid vault or rim fracture during reaming, central peg/base plate impaction, or peripheral screw insertion is less frequent than its humeral counterpart but carries important implications for baseplate fixation[30,39]. Pre-existing glenoid bone loss (long-standing CTA, rheumatoid erosion, avascular necrosis, or a prior failed implant) increases the risk of intraoperative fracture and complicate achievement of primary baseplate stability[1,39]. Management options mirror those developed for glenoid bone-loss reconstruction more broadly: Autograft or allograft bone grafting beneath an augmented or BIO baseplate, use of longer central posts/cages to engage residual vault bone. Moreover, staged bone grafting with delayed baseplate implantation may be favored in severe bony deficiency[1,39]. Occasionally, hemiarthroplasty without glenoid implantation is performed for non-reconstructable unstable glenoid, either as final construct or temporarily for later RSA[21]. Figure 3 illustrates a suggested algorithm for management of intraoperative humeral and glenoid-sided fractures.

Figure 3 Algorithm for recommended management of intraoperative fractures. BIO: Bony increase offset.

COMPONENT MALPOSITIONING

Excessive glenoid retroversion or superior inclination increases polyethylene edge-loading, accelerates baseplate loosening, and predisposes to inferior scapular notching, while humeral component malversion contributes to instability and impingement[1,16,40]. Because component position is the single most surgeon-controllable determinant of long-term implant survival, considerable recent effort has focused on enabling technologies: Preoperative three-dimensional computed tomography-based planning, patient-specific instrumentation (PSI), and intraoperative navigation or robotic-assisted preparation have each been shown to significantly improve the accuracy of glenoid baseplate version, inclination, and screw trajectory compared with standard instrumentation[40-44], although long-term data confirming a resultant reduction in loosening or notching rates are still maturing[44,45]. Given the cost and learning-curve considerations of these technologies, their selective use in cases with substantial glenoid bone loss or deformity currently represents the most defensible clinical application[40,43].

PHASE II (EARLY POSTOPERATIVE COMPLICATIONS)

The early postoperative window (conventionally the first three months), when most acute instability and virulent infections are present, is where mechanical intraoperative issues give way to biological and soft-tissue-tensioning problems[46,47].

Prosthetic instability

Instability remains among the most common and most disabling early complications of RSA, historically reported to account for approximately 4.7% of all complications, though its incidence has fallen with modern implants, with a systematic review of contemporary series reporting rates as low as 0.24% to 1%[6,48,49]. A recent systematic review of interventions and outcomes for RSA instability confirms that the great majority of episodes occur in the first six months. Roughly, half of these events are present within the first three months in an antero-superior direction driven by the arm being placed in the at-risk position of adduction, extension, and internal rotation[46,49,50]. Recognized risk factors include prior shoulder arthroplasty, inadequate soft-tissue tension from component malposition or an undersized glenosphere, subscapularis deficiency, axillary nerve or deltoid dysfunction, and bony deficiency compromising the deltoid wrap[51-54]. The role of subscapularis repair remains genuinely debated. Several meta-analyses report a significantly lower dislocation rate with repair[55,56]. Bethel et al[56] demonstrated 0.8% vs 4.2% in their recent meta-analysis favoring subscapularis repair. A randomized trial found comparable functional outcomes whether repair was performed. However, a later meta-analysis restricted to lateralized designs found dislocation rates were statistically similar with or without repair. That study suggested that lateralization itself may partly substitute for the stabilizing effect of subscapularis integrity[27,57,58]. Where the subscapularis tendon can be brought to an over-the-top position without excessive tension, repair is nonetheless favored by most contemporary series given its consistent association with reduced dislocation risk and no demonstrated functional loss[24,56,59,60]. Management follows a stepwise approach (Figure 4): Closed reduction under sedation is successful in roughly half of first-time dislocations and may be definitive, but recurrent instability, or instability associated with an identifiable structural cause (malposition, undersized glenosphere, subscapularis deficiency), generally requires revision - ranging from isolated liner exchange to a larger or more lateralized glenosphere, humeral component revision to correct version, or subscapularis reconstruction[49,61,62].

Figure 4 Algorithm for recommended management of postoperative instability. At-risk position: Adduction, extension, and internal rotation. RSA: Reverse shoulder arthroplasty.

Neurologic injury

Nerve injury after RSA is under-recognized when ascertained only clinically; intraoperative neuromonitoring studies suggest a true incidence considerably higher than the 1% to 4% typically reported by clinical examination alone[63-65]. The axillary nerve is most frequently implicated, owing to its intimate relationship to the inferior capsule and its vulnerability to traction during arm lengthening and lateralization, retractor placement, and inferior screw trajectory[65-67]. A correlation has been demonstrated between the degree of humeral distalization achieved and postoperative neurologic deficit, reinforcing that arm-lengthening beyond a certain threshold (most series suggest limiting lengthening to approximately 2 cm to 2.5 cm) carries a measurable neurologic cost[68]. Glenoid bone grafting for severe bone loss has separately been associated with a higher rate of early neuropraxia, presumably related to more extensive soft-tissue mobilization and retraction required for graft preparation and fixation[69]. Prevention centers on avoiding excessive arm lengthening, gentle and intermittent retraction (with periodic release), careful identification and protection of the axillary nerve at the inferior glenoid and conjoint tendon (the “tug test” remains a simple intraoperative safeguard), and a lower threshold for intraoperative neuromonitoring in complex revision or bone-grafting cases[69-71]. Most traction neurapraxias are transient, with most patients recovering motor function within six to twelve months with observation and physiotherapy; persistent deficits beyond this window warrant electrodiagnostic evaluation and consideration of peripheral nerve surgical consultation[63,65].

Hematoma

After RSA, there may be a significant dead area, especially if RC is significantly deficient, and there is a 1%-20% increased risk of hematoma formation. Hematomas may form because of inferior glenosphere placement, glenosphere medialization, or valgus orientation of the humeral component[72]. Intraoperative careful hemostasis layered wound closure, and drain can guard against its occurrence. A potential periprosthetic infection is associated with hematomas. Persistent discharge from the incision may be a sign of early sinus tract formation; therefore, debridement, irrigation, and hematoma evacuation may be necessary[21].

Early PJI

Acute infection is typically caused by virulent organisms such as Staphylococcus aureus and gram-negative bacilli (within the first four to six weeks of surgery). However, indolent low virulent Cutibacterium acnes infections often dominate the chronic late-presenting category[73-75] (discussed in phase III). Reported risk factors for periprosthetic infection after RSA include male sex, younger age, prior ipsilateral shoulder surgery, diabetes, obesity, immunosuppression, and hematoma formation[76-78]. In acute infections, the fibrous capsule is not yet matured around a well-fixed implant and often amenable in appropriately selected cases to irrigation and debridement with component retention alongside modular liner/humeral head exchange and prolonged culture-directed antibiotic therapy. Nonetheless, late infections that associate component loosening or biofilm-forming organisms most often require component removal[74,79,80]. Preventive strategies with reasonable evidentiary support include preoperative chlorhexidine skin preparation, alcohol and chlorhexidine surgical skin antisepsis (rather than iodine based), and appropriately timed perioperative antibiotic prophylaxis. Though, the marginal benefit of extended or targeted anti-Cutibacterium prophylaxis remains an area of active investigation[81-83].

Glenoid dissociation

The morse taper and the central locking screw are considered the two mechanical techniques utilized to secure the glenosphere to the baseplate in accordance with the RSA design[84]. There have been reports of glenosphere dissociation from the baseplate in a few RSA designs[85-87]. Soft tissue interposition, bony impingement, fluid in the female aspect of the assembly’s well, single-pole engagement of the glenosphere following its impaction under a slight angle, insufficient force applied to impact the glenosphere, and incomplete-seated glenosphere due to proud or cross-threaded screws within the baseplate are all common causes of Morse taper failure. The cold-weld between the glenosphere and the baseplate is reduced by incomplete glenosphere-seating, even if it is only 1 mm[88].

Preventing such complications requires careful consideration of the dimensions of the implant being used. The rim reamer is the most effective way to achieve proper glenosphere sitting over the baseplate. The rim reamer’s outer diameter is preferably greater than the glenosphere’s. Without it, the reamer’s ability to remove the possibly obstructive interfering bone would be reduced[89]. The primary technique for confirming full glenosphere seating is still vigorous glenosphere pull following impaction. Additionally, when examined on a post-operative radiograph, the glenosphere should be flush with the baseplate. Following surgery, patients need to exercise caution to prevent early impact loading[84].

PHASE III (LONG-TERM BIOLOGICAL AND FIXATION FAILURE)

Scapular notching

Scapular notching is defined by erosion of the inferior scapular neck by mechanical impingement of the humeral polyethylene cup against the scapula in adduction. It remains the most extensively studied radiographic finding after medialized RSA. Postoperative notching has been categorized based upon the defect size described by Nyffeler and Sirveaux grading system[17,26,90], as illustrated in Figure 5. In the same context, later studies documented notching also with rotational friction (frictional impingement). This repeated friction produces liner wear, and notching. The resultant released fragments out of friction induces inflammatory process and osteolysis[91].

Figure 5 Illustration image for the Nyffeler and Sirveaux grading system of scapular notching based upon the defect size. Grade 1 shows a defect within the inferior pillar of the scapular neck; grade 2 with defined bone defect under the level of the inferior screw within baseplate; grade 3 with defect extending over the inferior screw; and grade 4 with bone defect reaching the level of central peg.

Scapular notching is usually observed radiographically six months postoperatively[23]. Traditional medialized-design series reported notching in up to 50% to 70% of patients[15,28], correlating in some, though not all, series with inferior functional scores and progressive polyethylene wear or component loosening over time[16,90]. The shift to lateralized glenoid and/or humeral geometry has produced a marked reduction in notching, to well under 30% and, in several mid-term lateralized cohorts, to rates approaching zero[12,28,92]. Component inclination (a shift from 155° to 135° humeral NSA) and inferior glenosphere overhang/offset are the technically modifiable variables most consistently associated with reduced notching[16,92].

Clinical consequences of scapular notching are still controversial. Prosthetic instability, unexplained pain, and loosening might be related. Earlier investigations reported that neither the presence nor the size of notching impacted clinical scores[15,93]. On contrary, other reports demonstrated declined clinical outcome with notching[16,17]. There is no consensus on erosion progression. Grassi et al[94] and Werner et al[95] concluded that radiographic progress seemed to reach a plateau. Nonetheless, others reported erosion worsening with time.

Aseptic loosening (glenoid and humeral)

Aseptic glenoid baseplate loosening is a leading cause of mid- to long-term revision. A large single-prosthesis cohort study found aseptic glenoid loosening to be an important, if numerically modest, cause of failure, disproportionately associated with baseplate malposition, inadequate screw purchase, and pre-existing glenoid bone loss[39,96,97]. Revision for glenoid baseplate failure has been shown to yield reasonable functional outcomes when performed with adequate bone-stock reconstruction, though results are inferior to primary RSA[98,99].

Radiolucent lines next to the humeral stem indicate humeral loosening. Gilot et al[100] showed lines resembling to the Gruen et al’s classification for femoral stem loosening following total hip replacement. Figure 6 shows the differentiation of the bone adjacent to the humeral stem into eight zones. Additionally, the lines are categorized as < 1 mm, 1 mm to 1.5 mm, 1.51 mm to 2 mm, or > 2.01 mm based on their width. If there is a radiolucent line ≥ 2 mm in ≥ 3 zones, the humeral stem is radiographically at risk for clinical loosening[100]. Humeral component loosening often originates at one of two locations: Either within the cement mantle or whenever the metaphyseal neck unscrews from the humeral stem. In order to improve the interlock between the neck and stem, an integrated polyethylene bushing can be added to the screw threads[101].

Figure 6 Illustration image for humeral stem loosening defined by Gilot et al[100]. Bone adjacent to humeral stem is divided into 8 zones. Zones 1, 2, and 3 represent the lateral aspect of the stem at the proximal, middle, and distal thirds respectively. Zone 4 is the area around the distal stem tip. Zones 5, 6, 7, and 8 represent the medial portion of the stem from the distal, middle, proximal thirds, and base, respectively.

An analysis of 2342 cases found humeral component loosening to be associated with cementless fixation in poor-quality bone, stem undersizing, and inadequate metaphyseal fill[102]. Stress shielding of the proximal humerus (progressive bone resorption around a well-fixed stem due to altered load transfer) has been documented radiographically with certain cementless short-stem designs and, while frequently asymptomatic, may compromise future revision bone stock[1,103].

Acromial and scapular spine stress fracture

Acromial and scapular spine stress fracture (ASF) has emerged as the signature late complication of the lateralized-design era, occurring in approximately 1% to 5% of patients depending on the cohort and diagnostic rigor applied[19,104-106]. ASFs have been classified into different types by Crosby et al[107]: (1) Type I: Fracture of the anterior acromion; (2) Type II: Fracture of the posterior acromial body posterior to the acromioclavicular joint; and (3) Type III: Fracture of the scapular spine extending from the tip of the peripheral screw.

Biomechanical studies demonstrate that glenosphere lateralization and increased humeral distalization > 2.5 mm raise tensile strain across the acromion and scapular spine by increasing deltoid length and moment arm, providing mechanistic plausibility for the observed clinical association[18,29]. However, a 2024 multicenter analysis of a 135° inlay design found that up to 8 mm of glenoid-sided lateralization did not, in isolation, increase ASF risk, implicating excessive humeral distalization as the more proximate biomechanical driver[18]. Patient-specific risk factors consistently identified across systematic reviews include osteoporosis, female sex, rheumatoid arthritis, prior acromioplasty or subacromial decompression, and pre-existing acromial thinning[20,104-106,108], while a systematic review cautions that reporting of these risk factors remains methodologically inconsistent across the available literature, limiting definitive risk stratification[108]. Classification systems by Levy et al[109] and by Crosby et al[107] can guide treatment strategies. Most Levy type I and II fractures are managed successfully with activity modification and bracing, whereas type III fractures (through the scapular spine base, destabilizing the deltoid origin) more often require surgical fixation or are associated with a persistent functional deficit[107,109,110].

Chronic PJI

Late presenting infection is usually indolent and caused by low virulent biofilm forming organisms (Cutibacterium acnes, coagulase-negative staphylococci). Multiple extended cultures and adjunctive biomarkers are usually performed for diagnosis[74,111-113]. Management remains genuinely contested between single or two stage revision procedures. An earlier systematic review and meta-analysis found that single-stage revision was numerically more effective at infection eradication than two-stage revision. However, their findings were likely confounded by selection bias because two-stage protocols were disproportionately reserved for more virulent or drug-resistant organisms[114,115]. A separate meta-analysis restricted to shoulder PJI reported broadly comparable reinfection and reoperation rates between the two strategies[115,116]. Two-stage exchange, despite its greater burden on the patient (an interval antibiotic spacer period), has been shown in matched cohort analyses to achieve functional outcomes statistically comparable to primary RSA at intermediate-term follow-up[117,118]. From a health-economic perspective, a recent break-even cost analysis found that single-stage revision remains more cost-effective than two-stage revision provided the reinfection rate does not exceed approximately 54% to 62%, a threshold comfortably above reported clinical reinfection rates for single-stage protocols in most series[118]. Definitive randomized trials are still needed to resolve this debate. In the interim, single-stage revision might be applicable for identified antibiotic-sensitive, low-virulent organisms in a host with adequate soft tissue and bone stock. Whereas two-stage revision is generally preferred for polymicrobial infection, unidentified organisms, drug-resistant pathogens, or compromised soft-tissue/bone conditions[74,119,120].

COMPLICATIONS OF RSA FOR FRACTURE SEQUELAE, TUMOR RECONSTRUCTION, AND REVISION ARTHROPLASTY

The complication categories above are not indication specific. However, three reconstructive contexts consistently and significantly modify the magnitude of risk within each category. Fracture sequelae (malunion, nonunion, and post-traumatic avascular necrosis of the proximal humerus) present with distorted proximal humeral anatomy, tuberosity malposition or resorption, and frequently compromised RC and deltoid function, all of which elevate rates of instability, neurologic injury, and intraoperative fracture relative to primary RSA for atraumatic indications[121-123]. A recent National Joint Registry study spanning multiple indications found that trauma sequelae carried a significantly increased risk of revision compared with all other indications for RSA[9], and a systematic review of RSA performed for acute PHFs similarly reported higher complication and revision rates than those typically quoted for elective RSA[124]. Oncological reconstruction following proximal humeral tumor resection introduces unique challenges related to deltoid deficiency particularly when the deltoid origin or its innervation is sacrificed for oncological margins. In addition, the extensive bone loss may require endoprosthetic or allograft- prosthetic composite reconstruction, and soft-tissue reconstruction of the remaining RC remnant. Although the overall functional outcomes are favorable, the instability and dislocation rates are considerably higher than in non-oncological RSA[125-128]. Revision RSA carries the cumulative technical burden of compromised bone stock, retained hardware, scarred or absent RC and deltoid tissue, and higher intraoperative fracture and infection rates than primary surgery. The registry data confirms a stepwise increase in re-revision risk with each successive revision procedure[61,129,130]. These differential risks are summarized comparatively in Table 2.

Table 2 Differential complication risk in special reconstructive indications relative to primary reverse shoulder arthroplasty.

Context	Complications increased	Representative evidence
Acute PHFs/fracture sequelae	Instability, neurologic injury, intraoperative fracture, revision	National Joint Registry study: Significantly increased revision risk vs other indications; systematic review of RSA for PHF reporting higher complication/revision rates
Oncologic reconstruction	Instability/dislocation (especially with modular endoprostheses and deltoid deficiency)	Systematic reviews and meta-analyses of oncologic RSA reconstruction
Revision arthroplasty	Intraoperative fracture, infection, re-revision	Registry data showing stepwise increase in re-revision risk with successive revisions

PHFs: Proximal humeral fractures; RSA: Reverse shoulder arthroplasty.

DISCUSSION

The complication landscape of RSA has been transformed twice over the past two decades: First, by the sheer expansion of indications and case volume, which has broadened the population at risk to include more technically demanding reconstructive contexts; and second, by the deliberate biomechanical redesign of implants to reduce notching, which has, in a genuinely instructive example of engineering trade-offs in orthopedic implant design, reduced one failure mode while amplifying another (acromial and scapular spine stress fracture)[12,18,19,106]. This reinforces a broader methodological point: Much of the historical “15% to 24%” complication-rate literature predates the lateralized-design era and cannot be applied unmodified to counsel patients receiving contemporary implants[6,8]. Persistent inconsistency in complication definitions and classification systems across studies continues to hamper direct comparison and meta-analysis, and standardization, remains an important unmet need for instability and infection reporting specifically. Registry-level data, increasingly available from the National Joint Registry and the Australian Orthopedic Association National Joint Replacement Registry, offer a valuable, generalizable complement to single-institution case series, particularly for detecting differential risk by surgical indication (such as the elevated revision risk in trauma sequelae) that individual series are often underpowered to demonstrate[56]. Looking forward, enabling technologies (three-dimensional preoperative planning, PSI, and intraoperative navigation or robotics) offer a plausible pathway to reducing the component-malposition-driven subset of both early instability and late loosening/notching, though outcome data demonstrating this translation into reduced revision rates remain immature and merit dedicated long-term study[40,44,45]. Table 3 consolidates the incidence, associated risk factors, and preventive and management strategies discussed across the three phases into a single reference framework. A recommended timeline-based algorithm for prevention and management of potential RSA is summarized in Figure 7.

Table 3 Summary of reverse shoulder arthroplasty complications: Incidence, risk factors, prevention, and management.

Phase/complication	Incidence	Key risk factors	Prevention and management
Intraoperative complications
Intraoperative humeral fracture	1.4% overall; up to 13.6% in revision RSA	Osteoporosis, prior hardware/deformity, narrow canal, revision	Hand-reaming, avoid excessive stem impaction; ORIF or long-stem revision if displaced or stem loose
Intraoperative glenoid fracture	0.3%-0.4%	Long-standing CTA, rheumatoid erosion, AVN, revision	Autograft/allograft with augmented or BIO baseplate, longer central posts/cages, staged grafting in significant loss
Postoperative complications
Prosthetic instability	0.24%-4.7%	Prior arthroplasty, component malposition, undersized glenosphere, subscapularis deficiency, deltoid/axillary nerve dysfunction	Closed reduction for first-time events; revision (liner exchange, larger/Lateralized glenosphere, subscapularis reconstruction) for recurrent or structural instability
Neurologic injury	1%-4%	Excessive arm lengthening (> 2-2.5 cm), extensive retraction, glenoid bone grafting	Limit distalization, gentle intermittent retraction, tug test, low threshold for neuromonitoring in revisions
Hematoma	1%-20%	Inferior glenosphere placement, glenosphere medialization, valgus humeral component	Meticulous hemostasis, layered closure, drain use; early evacuation if infection suspected
Early PJI	1%-4%	Male sex, younger age, prior ipsilateral surgery, diabetes, obesity, immunosuppression, hematoma	DAIR with modular exchange and culture-directed antibiotics for acute, well-fixed implants
Glenoid dissociation	Uncommon	Incomplete glenosphere seating, soft-tissue interposition, inadequate impaction force	Rim-reamer sizing, verification by pull test and flush radiographic appearance
Long-term complications
Scapular notching	50%-70% (medialized RSA) vs < 30% (lateralized)	Medialized design, inferior baseplate tilt, glenosphere positioning	Inferior baseplate overhang/tilt, lateralized glenosphere/humeral geometry; largely observational once established
Aseptic loosening (glenoid/humeral)	0.7%-5.7%	Baseplate malposition, inadequate screw purchase, glenoid bone loss, cementless fixation in poor bone	Meticulous baseplate fixation technique, revision with bone-stock reconstruction when needed
Acromial/scapular spine stress fracture	1%-5%	Osteoporosis, female sex, rheumatoid arthritis, prior acromioplasty, excessive humeral distalization (> 2.5 mm)	Limit humeral distalization, activity modification/bracing for Levy I-II, surgical fixation for type III
Chronic PJI	1.4%	Cutibacterium acnes, coagulase-negative staphylococci, biofilm formation	Extended cultures and biomarkers; single- or two-stage revision individualized to organism virulence and host factors

RSA: Reverse shoulder arthroplasty; ORIF: Open reduction and internal fixation; CTA: Cuff tear arthropathy; AVN: Avascular necrosis; BIO: Bony increased offset; PJI: Prosthetic joint infection; DAIR: Debridement, antibiotics, and implant retention.

CONCLUSION

Despite an increasing understanding of the fundamental biomechanical principles underpinning RSA, complications continue to occur, even in the hands of experienced surgeons, and their character has shifted meaningfully as implant design has evolved from medialized to lateralized configurations. A structured, chronological approach can anticipate intraoperative mechanical risk, early postoperative instability/neurologic/infective risk, and late biological and fixation failure as three distinct but interrelated phases. Such approach can provide a practical clinical framework for prevention and evidence-based management. Careful patient selection, meticulous preoperative planning with particular attention to bone stock and lateralization strategy, precise intraoperative execution offers the surgeon the greatest opportunity to minimize complications and maximize the durability of RSA. A vigilant surveillance should be tailored to the specific risk profile of fracture-sequelae, oncological, and revision-arthroplasty patients to anticipate potential associated complications.

Figure 7 Timeline-based recommended algorithm for prevention and management of reverse shoulder arthroplasty complications. CT: Computed tomography; PSI: Patient specific instrumentation; PJI: Prosthetic joint infection; DAIR: Debridement, antibiotics, and implant retention; ORIF: Open reduction and internal fixation.