1 Introduction

Reverse total shoulder arthroplasty (RTSA) has emerged as a valuable surgical option across a wide spectrum of challenging shoulder conditions.^1–4 Its indications include advanced rotator cuff arthropathy, complex fracture sequelae with proximal humeral deformity and tuberosity migration, cases of osteolysis or nonunion, revision procedures in cuff-deficient shoulders, as well as selected tumor reconstructions.^5–9

Since its inception, RTSA has undergone continuous improvements in both prosthesis design and implantation techniques. Aseptic glenosphere loosening was one of the major failure mechanisms in early RTSA designs.¹⁰ Prosthesis failure was attributed to faulty design, particularly the lateral center of rotation relative to the implant-bone interface, which resulted in excessive bending and shear forces on the glenoid component.

In 1985, Paul Grammont designed a large, medially placed prosthesis without a neck, positioning the center of rotation at the glenoid prosthesis–bone interface.^2,11 This design was based on the theory that a constrained prosthesis with a lateral center of rotation increases the rotational moment arm, leading to loosening of the glenoid component. Medializing the center of rotation reduces bending and tensile forces at the implant–bone interface.

Although this approach significantly reduced rates of glenosphere loosening and failure, complications such as scapular notching, stress fracture of the acromion and scapular spine, instability,^12,13 and limited internal rotation (IR) and external rotation (ER) persisted.¹⁴ Scapular notching is defined as the impingement of the humeral implant against the inferior aspect of the scapular neck. Also referred to as an ‘Adduction Deficit,’ this usually occurs when the arm is in the resting position. These issues prompted further studies focusing on surgical implantation techniques and prosthesis design to address scapular notching, minimize impingement, and enhance range of motion (ROM).

There are a few clinical studies,¹⁵ cadaveric studies,^16,17 and computational studies related to prosthesis implantation and outcome evaluation or prediction.^18,19 However, none of these studies have taken a holistic approach by simultaneously considering variability in both humeral retroversion and glenosphere position. While previous computational modelling studies used CT scan data from cadavers, our study utilized preoperative CT data from patients who underwent RTSA.

The objective of this study was to generate computational 3D models capable of predicting the range of motion (ROM) up to the point of impingement, based on varying humeral retroversion and glenosphere tilt/translation, and to understand the optimum positioning of the prosthesis to maximize ROM without impingement.

2 Materials and methods

2.1

2.1 Equipment and softwares used

3D shoulder models were reconstructed based on the outlines of two-dimensional preoperative images of the scapula and humerus from patients who had undergone RTSA. A 3D model of the widely used Aequalis® Reversed Prosthesis (Tornier SAS, Montbonnot, France) was also generated from a Computer Aided Design (CAD) model (Fig. 1). Multiple sessions of virtual surgery were performed using the reconstructed models of the scapula, humerus, and prosthesis.

Close

Fig. 1 Generation of 3D models 3D models of scapula and humerus were generated from CT images of patients.

Export to PPT

2.2

2.2 Technique

Virtual RTSA surgery was performed with the following modifications to the 3D model.1.Modification of the scapula by trimming (diameter 36 mm) (Fig. 2

Close

Fig. 2 Glenoid trimmingGlenoid is modified for a 36mm diameter.

Export to PPT

).2.Insertion of the baseplate (diameter 29 mm) and the glenoid head (diameter 36 mm) (Fig. 3

Close

Fig. 3 Addition of glenoid componentsInsertion of the baseplate (diameter 29mm) and the glenoid head (diameter 36mm).

Export to PPT

).3.Resection of the humeral head and insertion of the stem (Fig. 4

Close

Fig. 4 Humerus preparationHumeral head is resected and the stem is inserted.

Export to PPT

).4.Assembly of the modified humerus, scapula, and glenoid components.

The final output after virtual surgery was the postoperative model in the neutral position. The human skeletal model was processed using visualization software (Amira®, VSG, USA), and the virtual surgery was performed using CAD software (Rapidform 2006, INUS Technology, South Korea).

To analyze bony impingement, shoulder movement was defined in terms of abduction, forward flexion, internal rotation, and external rotation. It was hypothesized that the humeral stem and the curved surface of the glenoid head remain fixed relative to each other during motion. Impingement between the scapula and humerus was quantified using the “Collision Detect” function provided by the Rapidform software, analyzing contact with the inferior glenoid, superior glenoid, coracoid, and acromion during scapular plane elevation ("scaption") (Fig. 5).

Close

Fig. 5 Impingement assessment using the Collision detect Function.

Export to PPT

3 Results

Out of the combinations tested, implantation of the glenosphere with 3 mm inferior translation and no inferior tilt, along with the humeral prosthesis positioned at 0° of retroversion, proved to be the most favourable combination in terms of maximum range of motion until impingement.

Implantation of the glenosphere with 3 mm inferior translation and 10° of inferior tilt showed improvements in adduction deficit, abduction, and rotation. However, forward elevation was highest in the group with no inferior translation and 0° of tilt. Increasing the retroversion of the humeral component resulted in decreased forward flexion and internal rotation.

3.1

3.1 Abduction

The T3 group consistently demonstrated the greatest abduction angles among all tested conditions, with values ranging from 78.3° at HRV (humeral retroversion) 0°–78.8° at HRV 20° (Fig. 6). This indicates that even with progressive increases in retroversion, the abduction capacity remained well-preserved in this group, with only minimal fluctuations across the tested range. The T3-tilt group also maintained comparatively high abduction values, varying narrowly between 75.9° and 76.2°, reflecting a stable performance that was only marginally lower than that of the T3 group. In contrast, the T0 and T0-tilt groups exhibited consistently lower abduction ranges. The T0 group showed an increase from 73.6° at HRV 0–75.2° at HRV 20, whereas the T0-tilt group ranged from 73° to 74.1° across the same HRV intervals. Although both of these groups demonstrated slight improvements with increasing retroversion, their overall abduction values remained notably below those of the T3 and T3-tilt groups. However, in regards of the superior glenoid, neutral humeral retroversion had more impingement, while 20° humeral retroversion showed lesser impingement. These findings suggest that abduction is better preserved in the T3 configuration, with tilt exerting only a limited influence on the overall motion profile.

Close

Fig. 6 Abduction ranges Figure shows abduction ranges across different values of inferior tilt, inferior displacement and humeral retroversion.

Export to PPT

3.2

3.2 Adduction deficit

The bar chart (Fig. 7) summarizes the effect of inferior translation, inferior tilt, and retroversion on adduction deficit. Inferior translation of 3 mm is associated with a marked reduction in adduction deficit compared with the baseline (approximately 1.5° vs. 6°). Conversely, inferior tilt shows no significant difference between 0° and 10°, with both conditions producing deficits close to 10°. Retroversion, however, demonstrates a progressive increase in adduction deficit with greater angular change, from approximately 9.7° at 0°–12.8° at 20°.

Close

Fig. 7 Adduction ranges Bar graph showing adduction deficit for different values of implant positions.

Export to PPT

The line graph (Fig. 8) illustrates the variation in adduction deficit across different humeral retroversion (HRV) angles (0°, 10°, and 20°) for four conditions: T0, T0-tilt, T3, and T3-tilt. Both T0 and T0-tilt groups demonstrate a progressive increase in adduction deficit with increasing HRV, with T0-tilt consistently showing slightly higher values compared to T0 (values rising from 9.7° at HRV 0–12.8° at HRV 20 for T0 and from 10.2° to 13.3° for T0-tilt). Similarly, T3 and T3-tilt follow an upward trajectory, although the absolute deficits are significantly lower than those observed in the T0 groups. The T3 and T3-tilt groups display lower baseline values (1.5° and 3.4° at HRV 0, respectively) but also show a consistent increase with humeral retroversion.

Close

Fig. 8 Line graph shows least impingement for the T3 group with no inferior tilt and 0° retroversion.

Export to PPT

Impingement between the inferior glenoid of the scapula and the humerus was greatest at a humeral retroversion of 20° and least at the neutral position. The group with a 3 mm inferior displacement of the glenosphere (T3, T0-tilt) also showed low rates of inferior impingement.

The lowest level of impingement was observed in the group with 3 mm inferior displacement without tilting (T3).

3.3

3.3 Forward elevation

The group with no inferior displacement, no tilt, and neutral humeral version demonstrated the maximum forward elevation/flexion (Fig. 9). The 3 mm inferior displacement groups (T3, T3-tilt) showed more impingement compared to the groups with no inferior displacement. Humeral retroversion of 20° also resulted in greater impingement than the neutral position.

Close

Fig. 9 Forward elevation range.

Export to PPT

3.4

3.4 Internal/external rotation

The range of internal rotation was greater in the 3 mm inferior displacement groups (T3, T3-tilt), with ROM decreasing with increasing retroversion of the humeral component (Fig. 10). Conversely, the range of external rotation was also greater in the 3 mm inferior displacement groups (T3, T3-tilt), but 20° of humeral retroversion demonstrated increased ROM compared to neutral (Fig. 11).

Close

Fig. 10 Range of internal rotation Increasing retroversion decreases the internal rotation.

Export to PPT

Close

Fig. 11 Range of external rotation Increasing retroversion increases the external rotation movement.

Export to PPT

4 Discussion

The positioning of the glenoid and the humeral components is crucial in determining the functional outcome of the patients, like the range of motion (ROM) of the shoulder, after reverse total shoulder arthroplasty. Impingement between bony structures and the prosthesis is an unwanted complication associated with component malpositioning.

From our study, we have concluded that a 3 mm inferior displacement of the glenosphere was found to be the most appropriate position to minimize scapular notching. In all ranges of motion except forward flexion, the 3 mm inferior displacement of the glenosphere was superior to the other position combinations. Humeral retroversion of 20° demonstrated better ROM in only abduction and external rotation in comparison to neutral version.

Several recent studies have made recommendations regarding the effect of glenoid placement on scapular notching and impingement-free abduction. Current evidence favors inferior placement of the glenoid component.^20–23

Similarly, recent studies have investigated the relationship between humeral component retroversion and impingement-free internal and external rotation, although the results have been contradictory.^24,25 The effect of glenosphere position on rotation has also been explored.²⁶

However, previous studies have not adopted a comprehensive approach, as they did not evaluate the combined effect of glenosphere position and humeral retroversion across different shoulder movements following RTSA. A more generalized three-dimensional model that can assess all movements —including adduction, abduction, forward elevation, and rotation—would further refine surgical techniques by providing valuable insights into optimal glenoid placement and humeral retroversion.

Virani et al.²⁷ in 2013, used a virtual three-dimensional saw-bone scapula/humerus model to study the effects of different combinations of RTSA components and baseplate positions on glenohumeral motion after RTSA. Their study focused primarily on implant-related factors, including type of humeral implant (inset/onset), glenosphere diameter (30, 36, and 42 mm), glenosphere center of rotation (COR) offset (0, 5, and 10 mm), humeral neck-shaft angle (130° and 150°), and humeral polyethylene socket offset (0, 5, and 10 mm). The only surgery-related factor considered was glenosphere placement (neutral/inferior).

The primary objective of our study was to determine the effect of surgery-related factors, such as glenosphere placement (neutral/3 mm inferior), glenosphere tilting (neutral/10-degree inferior tilt), and humeral version (0°, 10°, and 20° retroversion), on impingement during various shoulder movements. In our study, 3D scapulohumeral models were reconstructed from preoperative CT images of the scapula and humerus of patients who had undergone RTSA.

We agree with previous clinical and biomechanical studies that inferior prosthetic placement had the greatest effect on preventing adduction deficit.^22,28,29 However, excessive inferior placement may result in the inferior screw being placed below the scapula, "over-tension" of the deltoid, stretching of the neurovascular bundle, and an increased risk of fractures of the acromion or scapular spine.³⁰ Therefore, rather than placing the prosthesis too inferiorly, using a prosthetic design that allows the surgeon to create an inferior prosthetic overhang may help eliminate conflicts between the prosthesis and the inferior glenoid.³¹

In our study, inferior translation of the prosthesis also improved abduction ROM, as seen in previous reports. Superior impingement was at its lowest in the group, with only 3 mm inferior displacement without tilting (T3). Inferior placement of the prosthesis resulted in higher rotational movements, with both internal and external rotation being maximum in this group. However, very few studies have investigated how changes in glenosphere placement in RTSA affect internal and external rotation. Xinning et al.²⁶ and de Wilde et al.²² showed in their computer modelling studies that superior glenoid translation resulted in significant restrictions on rotation, while inferior placement led to an increase in both internal and external rotation. Virani et al.²⁷ also found that inferior glenoid placement was the most important factor leading to increased internal and external rotation ROM.

Boileau et al.³² and Valenti et al.³³ found that lateralization of the glenoid component improved rotation and reduced impingement, even though lateralization increases the bending moment arm and, thereby, shear forces at the bone-implant interface.

Our study did not show a reduction in impingement or improvement in ROM with inferior glenoid tilt. This contrasts with previous studies by Nyffeler et al.,³⁴ Kontaxis and Johnson et al.,³⁵ and Gutierrez et al.,^36,37 which demonstrated improved impingement-free ROM with inferior tilting of the glenoid component. Furthermore, inferior tilting may compromise prosthetic stability by reaming away the hard subchondral bone from the lower glenoid. In our study, the neutral positioning of the glenosphere showed a better forward flexion than inferior displacement and inferior tilt.

Biomechanical studies investigating the effect of humeral version on impingement after RTSA yield varying results. In general, decreased retroversion reduces external rotation and increases internal rotation, and vice versa. Grammont et al.³⁷ favoured 30° of humeral retroversion in their seminal description of RTSA, while Valenti et al.³³ used 20° of retroversion to align both the humeral and glenoid components along the same axis. In a cadaveric study, Stephenson et al.²⁵ suggested placing the humeral component between 20° and 40° of retroversion, while Berhouet et al.²⁴ found that maximum ROM was achieved with native retroversion, which averaged 17.5°. In contrast, Walch et al.³⁸ reported that less humeral component retroversion resulted in better outcomes in terms of activities of daily living (ADL), strength, constant score, and passive anterior elevation.

Gulotta et al., using a computer modelling study, concluded that increasing retroversion of the humeral component provided no benefit.¹⁸ They reported that decreasing retroversion increased internal rotation, and although external rotation at the side became limited, external rotation at 90° was unaffected. We agree with Walch et al. and Gulotta et al. that retroversion of the humeral component is not desirable for improved functional benefit. Our study reports a decrease in forward flexion and internal rotation with an increase in retroversion from 0 to 20°. An increase in adduction deficit was also associated with increased retroversion in our study, which can lead to scapular notching and early implant failure. Only external rotation was found to increase with increasing retroversion of the humeral component. Decreased external rotation with the arm at the side is not functionally debilitating, as patients can easily compensate by turning the body to the side. However, internal rotation limitation in adduction is more functionally debilitating, as it is important for performing activities of daily living (ADLs), such as perineal hygiene and reaching the back.³⁹

External rotation is functionally more important during shoulder abduction for activities of daily living (ADL).⁴⁰ Cadaveric studies show that at higher degrees of scaption, impingement becomes less pronounced,²⁵ and at 60° of scaption, models were able to rotate more than 90°. Increasing retroversion does not provide any mechanical advantage to the teres minor and posterior deltoid during arm elevation. Therefore, decreased retroversion will not limit external rotation during abduction but will improve internal rotation during adduction, which is more functionally important. Regarding stability, a French multicentre study found an association between neutral humeral version and the absence of dislocation episodes. Favre et al.⁴¹ also suggested limited retroversion for improved anterior stability. A recent study by Rhee et al.⁴² on the effects of component retroversion on functional outcomes showed no significant difference in terms of ROM and daily movements between 20° and 0° humeral component retroversion angles after RTSA. However, patients with 0° retroversion scored better on activities related to internal rotation.

Recent computer modelling studies, clinical studies on function, and studies on stability after RTSA all support placing the humeral component in less retroversion, which is in agreement with our results.

This study has certain limitations. First, the analysis was based on a two-dimensional model of the scapula and humerus derived from preoperative CT scans, without accounting for the role of muscles or surrounding soft tissues. Secondly, no statistical test can be applied to prove superiority, as the absolute values themselves serve as the proof. Thirdly, implant design variables were not examined, since these were kept constant in all simulations, including the use of a 36-mm glenosphere, absence of lateralization, and a 130° humeral neck-shaft angle.

5 Conclusion

Implantation of the glenosphere with 3 mm inferior translation, without inferior tilt, and with the humeral prosthesis at 0° of retroversion presents the most favourable combination in terms of range of motion without impingement.

Patient consent

No patient information or interventions were performed and patient data were used retrospectively.

Patient consents were taken for the same.

Contributions

Dr. Prince Shanavas Khan- Conceptualization, writing.

Dr. Yon-Sik Yoo- Conceptualization, writing.

Dr. Seong-Wook Jang- Conceptualization, writing.

Dr. Akhil Mathew Jacob- Writing – review & editing.

Dr. Ayyappan V. Nair- Conceptualization.

Dr. Nizaj Nasimudeen- Writing – review & editing.

Dr. Jimmy Joseph Meleppuram- Writing – review & editing.

Ethical statement

This study was conducted using anonymized preoperative CT data for the purpose of three-dimensional computational modelling. No interventions were performed on human participants or animals. The study adhered to the ethical principles outlined in the Declaration of Helsinki. Informed consent was waived as the study involved retrospective imaging data without any identifiable patient information.

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.