This study aimed to analyse moment arms and origin-to-insertion distance of the subscapularis and teres minor before and after reverse shoulder arthroplasty using a combined in-vitro/in-silico approach. Even though the functionally deficient infraspinatus may contribute to a loss of external rotation, the aim of this study was to investigate the effect RSA has on the intact muscles and their capability to perform rotational movements. Therefore the function of the infraspinatus was not specifically analysed in this study. This is the first study to characterise these properties after RSA. Knowledge of the functional properties of these muscles is of enormous importance for clinical practice and possible further improvement on prosthesis design or surgical technique.
Our pre-operative group consists of healthy shoulders, in which the humeral head is centered in the glenoid cavity. This might not be the case in shoulders with cuff tear arthropathy, but as the position of the humerus and therefore the center of rotation is highly variable in this pathology, we assumed this not practicable in terms of reproducibility. However, we assume in cases with a significantly cranialised humeral head the overall distalisation will be even more pronounced, leading to even more substantial changes in the joint's biomechanics.
The humeral component was implanted in ten degrees of retroversion in our entire specimens. Varying the humeral components' rotational alignment will likely have an impact on muscle tension. However in our opinion it is not an option to decrease muscle slackening as, for example, tensioning the posterior cuff will result in reduced tension of the anterior segments and vice versa. Also an increased retroversion might result in increased prosthetic impingent in neutral rotation or even increase the risk of prosthetic dislocation.
The methodology used is based on three-dimensional models derived from CT specimens' data. While CT scans allow reconstruction of the osseous anatomy with high precision, accuracy for identification of muscle origins and insertions was assumed not to be high enough. Therefore we marked muscle origins and insertions after preparation and visualisation using radio-opaque markers. Muscle wrapping was not included in this model, as it was considered negligible in the tested positions. Nonetheless we are aware of its possible impact to the overall value of our results. However, in our study we aimed to analyse the change of muscle properties rather than to obtain absolute values. The possible inaccuracy was therefore assumed acceptable. The pre-operative moment arm values calculated using this method are comparable to data from previous studies concerning normal shoulders [20, 21].
One of the major drawbacks of RSA is its lacking potential to improve active external and internal rotation. While in healthy shoulders external rotation is dependent on teres minor integrity, after RSA potential of external rotation remains small irrespective of its pre-operative status. Even in patients with only mild fatty degeneration preoperatively, the gain in active external rotation remains small. Patients with higher grade fatty infiltration pre-operatively, might even experience a loss in external rotation. While Boileau et al.  propose several reasons, such as prosthesis design and altered biomechanical properties of the deltoid, as being responsible for this, postoperative changes to the teres minor's rotational moment arms and origin-to-insertion distance, as shown in our study might be another, important contributing factor. Rotational moment arms are significantly smaller for all but the 15 degrees position, even though a corresponding trend in this position can be seen as well. Additionally muscle slackening might further reduce its efficiency, as origin-to-insertion distance is significantly smaller, especially in the 15 degrees position, reaching up to 20 mm for the distal segment.
Accordingly internal rotation, which in healthy shoulders depends on intact subscapularis function, often is compromised after RSA as well. The subscapularis muscle tendon unit is the main internal rotator and contributes considerably to active stabilisation of the glenohumeral joint. In this study the two more cranial segments had significant smaller rotational moment arms after RSA, while no difference could be seen for the distal segment. No definite rational can be given to explain this difference. Further mathematical analysis might therefore be necessary.
While failed or non-performed reconstruction of the subscapularis has shown to have an influence on clinical outcome in anatomical shoulder arthroplasty, no difference was seen after RSA at this stage. Even though Edwards et al.  identified impaired subscapularis integrity at the time of surgery as the most important risk factor for dislocations in shoulders where reconstruction was impossible due to insufficient proximal humerus bone stock, no higher risk was seen in patients with cuff tear arthropathy as aetiology. Unfavourable biomechanical properties after RSA, as shown in this study, with a decreased moment arm in conjunction with the decreased muscle tension might impede better results, no matter if the subscapularis is reconstructed or not. On the other hand, its integrity might have been irreversibly impaired pre-operatively or secondary to the surgical approach.
Differences of the origin-to-insertion distances were most pronounced for the cranial segments in the 15 and 30 degrees abduction positions for both muscles. With increasing abduction this difference decreases and for some segments and positions no significant difference can be seen. We assume that with implantation of the RSA and distalisation of the humerus an increased distance of the tendon insertions to the rotational center arises. This results in a more eccentric motion of these landmarks and might explains the decrease of the origin-to-insertion distance with increasing abduction.
In both muscles some segments had positive abduction moment arms preoperatively, which in healthy shoulders is essential for their function as dynamic stabilisers of the shoulder joint. The loss of this function will lead to a smaller joint compression force and as a result increase subluxation forces. These increased forces might abet glenoid loosening and instability. No beneficial effect can be seen for the increased postoperative adduction moment arms as adduction is usually not impaired in patients with cuff arthropathy, neither pre- nor postoperatively.
Scapular notching is one major complication in reverse shoulder arthroplasty. Mechanical impingement as well as secondary bone erosion due to polyethylene wear is believed to contribute to this phenomenon. In our study, inferior impingement between the humeral component and the scapular neck was only observed in the zero degree reference position, which, however, is not the neutral thoraco-humeral position, but rather an adduction position which is not of high clinical relevance. Even though scapular notching was not the specific focus of this study, these findings are in agreement with the observations of other authors on this subject.