1 Introduction

In light of the reliable clinical results by treating glenohumeral arthritis, cuff-tear arthropathy and non-reconstructable fractures of the proximal humerus with reverse shoulder arthroplasty, ongoing research on the matter has been brought forward. But in spite of its great popularity, reverse shoulder arthroplasty is associated with a high rate of complications, including scapular notching, baseplate failure, periprosthetic fractures, instability, nerve lesions and deltoid-insufficiency.^1,2 Besides others, sufficient function of the deltoid muscle by finding the optimal position of the baseplate has been proven to be a crucial factor for obtaining good clinical results.^3–5 In spite of Grammont's groundbreaking concept to shift the center of rotation inferiorly and medially with respect to the native joint, with increasing experience lateralization of the center of rotation has been suggested as a reliable option to improve range of motion (ROM) and limit scapular notching.^6–10 Lateralization can be achieved by extended glenosphere components or bone spacers.^6,11 Although clinical results suggest improvements in external and internal rotation, the effect of lateralization on joint motion in reverse shoulder arthroplasty is discussed controversially.^12–14

Therefore, the present study aims to analyze the effect of glenoid lateralization on active range of motion, using a dynamic test setup in a cadaveric model with pre-loading the muscles of the rotator cuff and the deltoid muscle. According to recent studies, we hypothesized that increased lateralization will improve ROM, while excess lateralization might lead to a loss of function.

2 Materials and methods

2.1

2.1 Specimens

Reverse shoulder arthroplasty was performed in 7 fresh frozen cadaveric shoulder specimens by a single surgeon (NO): 4 left and 3 right specimens (4 female and 3 male) with a mean age at death of 71 years (min. 52; max. 87) were available. The specimens included the shoulder joint with scapula, clavicle, and the proximal half of the humerus, as well as the intact soft tissue envelope.

Using fluoroscopy, shoulders were checked for previous, relevant bony injuries, previous surgery or bony glenoid defects (Fig. 1). Prior to testing, the specimens were thawed at room temperature.

Close

Fig. 1 Using fluoroscopy shoulders with a fracture of the structures relevant to our study, previous surgery or bony glenoid defects were excluded; anterior-posterior and axial view of shoulder.

Export to PPT

2.2

2.2 Reverse shoulder arthroplasty

All reverse shoulder prostheses were implanted through a delto-pectoral approach by a single specialized surgeon (NO). In each cadaveric shoulder the Delta Xtend reverse shoulder prosthesis (Cementless Modular Humeral Implant, Fa. Depuy/Synthes) was implanted. We obtained the subscapularis (SSC), infraspinatus (IF) and teres minor (TMI) tendons; the supraspinatus tendon was sectioned.

The humeral head was osteotomized in 10° of retroversion. The humeral canal was then reamed, and a standard-length humeral stem, 10 mm or 12 mm, was implanted press-fit.

To prepare the metaglene, the glenoid was reamed and a baseplate was fixed into position with two 3.5-mm non-locking screws. A bi-cortical screw placement was achieved in each setup. A 38 mm glenosphere with a neutral offset was placed in all cases. For the increments of lateralization we used hand-fabricated PMMA-spacers (Fig. 2). Spacers were available in thickness of +5 mm (subgroup II) and +10 mm (subgroup III). Subgroup I was measured without lateralization. The metaglene was removed for each trial to add the spacer, the same holes were used and longer screws were needed.

Close

Fig. 2 Circular PMMA-spacers, manufactured to fit under the metaglene, without relevant overhang; thickness of +5 mm and +10 mm.

Export to PPT

In all cases we used a 38 + 3 mm Inlay. The origins of the SSC, IF, TMI and deltoid (each segment: pars clavicularis, acromialis and spinalis) were sectioned proximally and each insertion was augmented with fingertips and No. 2 FiberWire (Arthrex Inc., Naples, FL, USA). The FiberWires were led close to their bony insertions to replicate the force vector optimally (Fig. 3), and were connected backwards.

Close

Fig. 3 Fiber wires were led close to their bony insertions to replicate the force vector optimally; DA pars acromialis, DV pars clavicularis, DS pars spinalis, TM teres minor, IS/ISF infraspinatus, SSC/SCP subscapularis.

Export to PPT

3 Results

3.4

3.4 Group III

Abduction (pars acromialis) was by using a maximal load (+4 kg) 22° (14°, 22°; 61°). External rotation (pars spinalis) was 4° (-; 4°; 26°) by a maximal load and 2° (2°; 3°; 12°) by using a load of 2 kg. The internal rotation (pars clavicularis) was 18° (12°; 18°; 36°) by a maximal load and 15° (5°; 15°; 26°) by a load of 3 kg. The external rotation was 54° (25°; 54°; 69°) caused by TMI and 34° (10°; 34°; 61°) caused by IS at maximum load.

The internal rotation (SSC) was 17° (10°; 17°; 33°) by using a load of 1,5 kg.

A higher mean external rotation by IS and TMI can be reached in subgroup II by using a less load compared to subgroup I (Fig. 4a).

Close

Fig. 4a better external rotation (teres minor, TM; infraspinatus, ISF) in subgroup II with a load of 2 kg compared to subgroup I with a load of 4 kg; x-axis subgroup, y-axis range of motion in degrees (external rotation).

Export to PPT

Higher internal rotation (SSC) can be reached with a less load in subgroup II compared to subgroup I.

When loading the TMI, ISF and SCP with more than 1 kg, dislocations can be reported. Table 2 summarizes the rates of dislocations depending on the used load.

Table 2 Rates of dislocations depending on the used load, not reported by RoM; Subgroups I, II, III with increments of lateralization, DV pars clavicularis, DS pars spinalis, TM teres minor, IS infraspinatus, SSC subscapularis.

kg	1	1,5	2	2,5	3	3,5	4	total
DS (I)	0	1	2	0	0	0	1	4
DC (I)	0	0	1	1	1	1	1	5
TM (I)	0	0	2	0	0	0	1	3
IS (I)	1	2	0	0	0	0	0	3
SSC (I)	0	0	3	0	0	1	0	4	19
DS (II)	1	0	1	0	2	1	0	5
DC (II)	1	0	0	1	1	0	1	4
TM (II)	1	2	0	0	0	0	0	3
IS (II)	0	2	1	0	0	0	0	3
SSC (II)	0	1	1	0	1	1	0	4	19
DS (III)	1	0	0	1	0	1	1	4
DC (III)	0	0	1	1	0	1	0	3
TM (III)	0	1	1	0	1	0	0	3
IS (III)	1	1	0	1	0	0	0	3
SSC (III)	0	1	1	1	1	0	0	4	17

3.5

3.5 Statistical analysis

The Wilcoxon signed rank test was used to evaluate the difference between the groups. The mean abduction was higher in group II, compared to group I and group III (Fig. 4b), but the results were statistically not significant (p = .063 and p = 0.091 respectively). Moreover, for internal rotation (SCP/pars clavicularis) and external rotation (TMI/ISF/pars spinalis) no significant difference between the groups can be reported (p > .05).

Close

Fig. 4b Boxplot abduction in degrees with a maximal load (4 kg) of pars acromialis (DA); x-axis subgroup I/II/III, y-axis abduction in degrees.

Export to PPT

4 Discussion

The purpose of this study was to analyze the effect of lateralization of the glenosphere in RSA on ROM (abduction, internal and external rotation). The present study has shown that lateralization could influence ROM after reverse total shoulder arthroplasty in an in-vitro setting. In comparison to a +0 mm lateralization, + 5 mm lateralization showed higher ROM with less load, while +10 mm lateralization achieved less ROM.

We have started our experimental setup with 5 specimens. However, the results have not shown a significant difference. So, in order to ensure this, the number of specimens needed for the verification of the hypothesis had to be calculated with the aid of a power analysis. For the verification of the hypothesis 2 more specimens were needed. Nevertheless, we were not able to reach a level of significance.

While the configuration of reverse shoulder prosthesis can be altered by many design parameters, in the present study we focused on glenosphere offset, as we believe this to be highly influential on soft-tissue tensioning, joint load and therewith ROM. A few cadaver studies elaborated specifically on the effects of glenosphere diameter and lateralization on joint stability and RoM.^12,15–17 Especially the influence of lateralization on stability is discussed controversially. Ferle, Pastor and Smith recently published a biomechanical investigation on the topic, using a similar cadaver setup with a robotic shoulder simulator, including the effect of humeral component characteristics and glenosphere lateralization on stability after RSA.¹⁶ They found significantly higher construct stability when the glenosphere was lateralized by 6 or 9 mm. However, influence of lateralization on ROM was not measured. Our results have not shown a significant difference of lateralized RSA on joint stability. Henninger et al. have shown increased abduction after COR lateralization, while no influence on adduction and external rotation could be observed.¹² Greiner et al. observed an increased rotational moment arm for TMI and SCP by using a 3D shoulder model in lateralized RSA in comparison to non-lateralized RSA.⁹ Our results have shown that a higher external and internal rotation can be reached with less load in lateralized RSA.

Their setup allows precise measurement and a high level of repeatability by using a robotic shoulder system. We aimed for an alternative experimental setup. It is reasonable to assume muscle force as a driving factor in real live range of motion. To our knowledge, this is the first biomechanical study to apply such a technique in a cadaveric in-vitro model of shoulder specimens.

While biomechanical studies on the subject showed measurable effects on joint load and improvement of RoM,^9,16,18 a systematic review of clinical studies by Helmkamp et al. reported improved external rotation and reduced scapular notching in the group of implants with lateralized COR.¹⁹ These results of Helmkamp's group basically confirmed the reported data from the systematic review of Lawrence et al., in 2016²⁰, who also reported better external rotation with less scapular notching. Our results confirmed better external and internal rotation. So, a higher external rotation could be reached by using a load of 2 kg in subgroup II in comparison to subgroup I using a load of 4 kg. In lateralized RSA less load could lead to higher external rotation. Notably, both reviews found a higher rate of glenoid loosening in the lateralized group,^19,20 which is in accordance with basic science work from Yang et al.²¹ and Denard et al.²² Yang et al. investigated the effect of glenosphere diameter and eccentricity on shear forces acting on the baseplate in RSA.²¹ Their 2-D finite element study convincingly showed an increase of stress at the base of the glenoid, at the inferior screw, with increasing glenoid eccentricity and offset. In a 3-D finite element setup, Denard et al. investigated stress and displacement at the glenoid in RSA, depending on glenosphere diameter and lateralization.²² Notably, their setup differentiated in bony lateralization and lateralization directly via the glenoid component. They found a significant increase of stress and displacement at the glenoid with increasing diameter of the glenosphere, and with lateralization, while the stress increased stronger with bony lateralization than with prosthetic. These studies support the clinical findings of the reports by Lawrence et al.²⁰ and Helmkamp et al.,¹⁹ as increased mechanical stress is a reasonable explanation of loosening. Zumstein's group reported a significantly higher rate of aseptic loosening of the glenoid component in a lateralized implant, compared to a medialized COR group.² Our results showed an improvement of abduction, internal and external rotation using +5 mm lateralization. However, + 10 mm lateralization decreased RoM. According to recent studies, lateralization could improve RoM, but an increased force is needed to abduct the arm.^11,17,23 Pronounced deltoid fatigue and acromial stress fractures could be the result.^23,24 The benefit of lateralization is to achieve the best balance of soft-tissue tensioning and joint load. Using a 5 mm lateralization was our best balance regarding to ROM.

The present study has some limitations. Firstly, we only performed reverse shoulder arthroplasty in 7 specimens because of the complex biomechanical setup and the general cost of cadaveric specimens. Hence, the values measured for each specimen did not show a wide variance. A preliminary study with 5 specimens comparing the impact of 0 mm offset and 5 mm offset on the range of motion of the acromial deltoid showed that 7 specimens are required in order to detect significant differences in the range of motion with a statistical power of 0.8. Although an effort was made to create a nearly anatomical setup, an incorrect estimation of the tension of the deltoid was inevitable. As we were aware of this potential measurement error, we recalibrated each time before measuring. The muscle bellies were not loaded dynamically. Therefore, comparability of the results with native biomechanics is not reliably possible.

To our knowledge, this is the first study to analyze the influence of lateralization on real-time range of motion. Our results hypothesized that lateralization could improve RoM, while excessive lateralization decreases RoM.

5 Conclusion

According to our results lateralization of the glenosphere may improve the range of motion in reverse shoulder arthroplasty. Using lateralization in reverse shoulder arthroplasty an offset of 5 mm seems to be favorable. Nevertheless, a detailed CT-analysis is essential for discussing a useful lateralization in reverse shoulder arthroplasty.

Disclaimer

No grant or grant support; All procedures performed in our study involving human cadavers were in accordance with the ethical standards of the institutional research committee (Ethical Committee of the Medical Faculty of the University of Cologne – VT 19–1540).