Primary osteoarthritis of the shoulder: are there differences in scapular morphology, orientation, and rotator cuff action lines between patients and healthy controls?

Abstract

Background

There is currently no in depth description of glenoid vault morphology, the rotator cuff action lines, and scapulothoracic orientation in patients with shoulder osteoarthritis (OA). Therefore, the goal of this study was to provide a quantitative analysis of these parameters in Walch types A, B, and C, along with rotator cuff action lines, scapulothoracic orientation, and humeral head centering compared to healthy controls.

Methods

Patients who were treated for primary shoulder OA in our hospital between 2010 and 2018 were included in this retrospective case-control study. The cases were categorized into type A, B, and C according to their glenoid morphology on computed tomography using the modified Walch classification and compared with a healthy control group in a matched-pair analysis (matching by age, gender, and affected side). The glenoid version, glenoid inclination, glenohumeral and scapulohumeral head centering, neck angle, glenoid and humeral offset, subscapularis (SSC) and infraspinatus (ISP) tendon traction vectors, and an overall rotator cuff (RC) vector were measured in a standardized axial plane. The protraction, internal rotation, upward rotation, translation, and tilt of the scapula were also measured in three-dimensional reconstructions.

Results

A total of 59 shoulders in 47 patients were identified with the following distribution of Walch glenoid types: 24 type A, 30 type B, and 5 type C glenoids. Type A glenoids showed no differences compared to their control group except significantly higher SSC angle, higher resultant RC vector, and lower scapular tilt. Type B glenoids had higher glenoid version, lower glenoid inclination, higher SSC angle, lower ISP angle, higher resultant RC vector, and more posterior humeral offset in comparison with their control group. Scapulothoracic orientation measurements for type B glenoids indicated significantly reduced scapular internal and upward rotation, and lower scapular tilt compared to controls. Type C glenoids showed significantly higher glenoid version, lower glenoid inclination, higher posterior humeral offset, and lower ISP angle with no significant change in the resultant RC vector in comparison with their control group.

Conclusion

Patients with primary OA and glenoid type A barely show differences in scapular morphology and scapulathoracic orientation compared to a healthy control group. In contrast, patients with glenoid type B or C showed significant differences of scapulohumeral centering in glenoid version, humeral offset, and glenoid inclination compared to controls. Furthermore, despite the changes of SSP and ISP angles in type B glenoids, the humeral head stayed centered to the glenoid fossa.

Primary and secondary glenohumeral osteoarthritis (OA) can be distinguished based on whether a specific cause for the OA is identifiable. Although predisposing exogenous factors such as inflammation, instability, previous surgery, or trauma can lead to secondary OA,¹⁵ primary OA cannot be attributed to a single specific cause but rather has a multifactorial pathoetiology, with age playing an important role.⁸

A further distinction can be made between concentric and eccentric OA. The Walch classification offers a qualitative description of wear patterns and the shape of the glenoid articular surface, thus providing information on whether a patient has concentric or eccentric OA changes.^5,49 This distinction is highly relevant for the choice of appropriate treatment, as inferior outcomes and higher complication rates have been described for anatomical shoulder arthroplasty in patients with eccentric OA and severe posterior glenoid erosion.⁵⁰ Several different treatment options to address severe posterior glenoid erosion in eccentric OA have been proposed, including noncorrective arthroplasty,⁴³ the ream and run procedure,²⁶ high side reaming and anatomical glenoid component implantation,^14,36 posterior bone graft augmentation of the anatomical glenoid component,^33,41 augmented anatomical glenoid components,^12,17,19,45 and reverse total shoulder arthroplasty.^28,29 Since no clear superiority of one concept over the others has been shown and eccentric OA remains a treatment challenge, it is crucial to analyze the pathomorphological discrepancies of concentric and eccentric OA more extensively. Although differences in glenoid articular surface shape and erosion,^21,22 humeral retrotorsion,³⁹ acromion shape,⁴ as well as rotator cuff (RC) muscle volume² and atrophy¹¹ have been described, there is currently no information on the shape of the scapular neck, RC action lines, or scapulothoracic orientation in patients with concentric and eccentric OA. Therefore, the goal of this study was to provide a quantitative comparison of scapular morphology and scapulothoracic orientation differences in concentric and eccentric OA along with their implications on the RC action lines and the resultant humeral head centering.

Methods

Study population

A retrospective case-control study was conducted, analyzing patients with primary OA of the shoulder, who were treated in our institution between 2010 and 2018. Exclusion criteria were age over 70 years, as well as previous shoulder surgery, including RC repair and instability surgery. These patients were then matched according to their age (within 5 years), sex, and affected side with patients from our institutional radiology database who had received computed tomography (CT) scans meeting the following inclusion criteria:

• 1)Supine position with the arms at the side
• 2)Complete depiction of the scapula, humeral head, and spine
• 3)Centered humeral head with no signs of glenohumeral arthritis

Patients with visual pathologies of the upper extremities and those with the presence of surgical hardware were excluded. Cases were graded according to glenoid morphology on two-dimensional (2D) CT scans using the modified Walch classification, which consists of six subtypes (A1, A2, B1, B2, B3, and C), through a consensus reading by 3 senior surgeons.⁵

Image selection for measurements

A standardized axial imaging plane was created for all patients using a multiplanar reconstruction, as previously described.^1,32 The standardized axial imaging plane was defined as the perpendicular imaging plane to the long axis of the glenoid that passes through the center of the best-fit circle drawn on the en face view of the glenoid.^1,31 Furthermore, CT data were rendered into three-dimensional (3D) models in patients with depiction of the complete scapula, humeral head, and vertebral bodies between T1 and T10.

Image measurements

All measurements, as well as 3D models, were independently performed once by 2 senior surgeons using Visage software (version 7.1; Visage Imaging GmbH, Berlin, Germany) for the purpose of determining interrater reliability.

2D measurements

Glenoid version was measured according to the method of Friedman et al¹³ and relative to the scapular blade axis as described by Hoenecke et al.²⁰ The glenoid axis was defined as the tangent between the anterior and posterior rims of the glenoid surface, and the scapular blade axis as the line of best fit between the medial border of the scapula and the beginning of the glenoid vault.²⁰ Glenoid inclination was measured on coronal CT images following the method outlined by Maurer et al.²⁷ The coronal slice that most accurately captured the floor of the supraspinatus fossa was chosen for these measurements. A line was drawn connecting the superior and inferior edges of the glenoid, referred to as the “glenoid face vertical line”. Another line was traced along the floor of the supraspinatus fossa, termed the “supraspinatus fossa line”. The angle between these two lines was then measured to determine the glenoid inclination angle. Glenohumeral and scapulohumeral subluxation indexes as described by Walch et al according to the glenoid plane at the level of the middle of the glenoid were used to calculate humeral head subluxation.^1,10,48 Furthermore, the neck angle and glenoid offset were calculated to measure the displacement of the glenoid vault in relation to the scapular blade axis, as previously described.¹

The line of pull was drawn for subscapularis (SSC) and infraspinatus (ISP), and the angle of pull in relation to glenoid surface was measured for both. First, the scapular blade axis was drawn, then the best-fit circle of the humeral head was placed, and a line perpendicular to scapular blade axis was drawn through center of the best-fit circle. The two points where this line intersects the best-fit circle were defined as the ISP insertion point posteriorly and the SSC insertion point anteriorly. From the posterior point, two lines were drawn: the first line to lateral border of the ISP origin in the ISP fossa and the second to the medial point of the ISP origin in the ISP fossa. Then, the bisector of these two lines was marked as the resultant ISP action line. The posterior-directed angle between resultant ISP action line and the glenoid surface line was measured and defined as the ISP angle. For the SSC angle, the starting point was the previously described SSC insertion point. Since 2 lines as described for the ISP cannot be drawn for the SSC without crossing the bony glenoid, a single line was drawn to the anterior glenoid rim and was defined as the SSC action line. The posterior-directed angle between this line and the glenoid surface line was the SSC angle. The resultant RC action line was determined as the mean angle of the ISP and SSC angles. The measurement of RC action lines is presented in Figure 1.¹⁶

Figure 1.

The measurement of rotator cuff action lines depicted on axial computed tomographic scans. (A) Infraspinatus (ISP) action line (green), (B) subscapularis (SSC) action line (yellow), and (C) resultant rotator cuff action line (red). Reprinted from Journal of Shoulder and Elbow Surgery, 34/3, Adrian Góralczyk, Doruk Akgun, Paul Siegert, Jonas Pawelke, Krzysztof Hermanowicz, Matthias Flury, Beat R. Simmen, Markus Scheibel, and Philipp Moroder, Humeral rotation osteotomy is not associated with glenohumeral and scapulohumeral decentering at long-term follow-up, 7, Copyright (2025), with permission from Journal of Shoulder and Elbow Surgery Board of Trustees.

In addition, glenoid vault tapering from lateral to medial was performed similarly to previously published methods.^1,31 The vault extent was measured as the distance perpendicular to the scapular blade axis from the anterior and posterior bony edges of the glenoid. The measurements were performed in 5 mm steps for a total of 13 times (Fig. 2). Furthermore, humeral offset was measured as the perpendicular distance to the scapular blade axis from the midpoint of the humeral head.¹

Figure 2.

The tapering of the glenoid vault from lateral to medial and the humeral offset (HO). SAx, scapular blade axis. Reprinted from Journal of Shoulder and Elbow Surgery, 30/4, Doruk Akgun, Paul Siegert, Victor Danzinger, Fabian Plachel, Marvin Minkus, Kathi Thiele, and Philipp Moroder, Glenoid vault and humeral head alignment in relation to the scapular blade axis in young patients with pre-osteoarthritic static posterior subluxation of the humeral head, 756-762, Copyright (2021), with permission from Journal of Shoulder and Elbow Surgery Board of Trustees.

3D measurements of scapulothoracic orientation

CT data were rendered into 3D models in 35 pathologic shoulders, depicting the complete scapula, humeral head, and vertebral bodies between T1 and T10. As previously described by Moroder et al,³⁰ protraction, scapular internal rotation, upward rotation, translation, and tilt were measured in the 3D models. Protraction was measured as the angle between the sagittal vertebral axis and a line from the glenoid to the center of vertebral body T1 in the axial plane. Scapular internal rotation was defined as the angle between the line from the glenoid to the root of the scapular spine in the transverse plane and a perpendicular line to the sagittal vertebral axis through the glenoid. Scapular upward rotation was determined as the angle between the longitudinal vertebral axis and the line from the root of the scapular spine to the inferior scapular angle in the frontal plane. Scapular translation was measured as the angle between the longitudinal vertebral axis and the line from the tip of the spinous process of the T1 vertebra to the glenoid in the frontal plane. Scapular tilt was determined as the angle between a perpendicular line to the upper baseplate of the vertebral body T1 and the line from the medial root of the scapular spine to the inferior scapular angle in the sagittal plane.³⁰

Statistical analysis

The paired samples t-test (for parametric distribution) or Wilcoxon test (for non-parametric distribution) were used to compare parameters between the matched case and control groups. Case groups were compared with each other using analysis of variance with Bonferroni correction for multiple comparisons (for parametric distribution) or the Kruskal–Wallis test with pairwise comparisons (for non-parametric distribution). Interrater agreement was assessed to determine the reliability between the different raters. All data were analyzed using Fleiss' K correlation coefficient to measure the level of interrater reliability. The Landis and Koch criteria were used for interpretation: values of 0.00-0.20 corresponded to slight agreement, 0.21-0.40 to fair agreement, 0.41-0.60 to moderate agreement, 0.61-0.80 to substantial agreement, and 0.81-1.00 to almost perfect agreement.²³ Statistical analysis was performed using SPSS (version 29; IBM Corp., Armonk, NY, USA). The standard significance criterion of $\alpha$ = 0.05 was used.

Results

For the case group, a total of 59 shoulders in 47 patients were identified with the following distribution of Walch glenoid types: 24 type A, 30 type B, and 5 type C glenoids. Twenty-one patients (45%) were female, with a mean age of 55.8 years (min = 32, max = 70, standard deviation = 8.35). Fifty-nine shoulders in 50 healthy patients were identified as the control group. In the control group, 18 patients (36%) were female, with a mean age of 56.1 years (min = 30, max = 69, standard deviation = 8.43). No significant difference was found between the case and control groups in terms of age, sex, and side (Table I). According to intraclass correlation coefficient (ICC) calculations, every measurement taken was in almost perfect agreement. The ICC_mean for interrater reliability had a value of 0.84. All measurements showed moderate to almost perfect agreement between the two raters. ICCs for every measurement are summarized in Table II.

Table I.

Demographic characteristics of OA and control group.

	OA group (n = 59)	Control group (n = 59)	P value
Age (yr)	55.8 ± 8.4	56.1 ± 8.4	.83
Sex	26 female, 33 male	23 female, 36 male	.57
Side	31 right, 28 left	31 right, 28 left	1.0

OA, osteoarthritis.

Table II.

Calculated ICC for all measurement parameters.

	ICC	95% confidence interval		Reliability
		Lower bound	Upper bound
Glenoid version according to the method of Friedman	0.970	0.957	0.979	Almost perfect
Glenoid version relative to scapular blade axis	0.646	0.491	0.753	Moderate
Glenoid inclination	0.928	0.897	0.950	Almost perfect
Glenohumeral subluxation index	0.743	0.626	0.822	Moderate
Scapulohumeral subluxation index	0.933	0.892	0.957	Almost perfect
Neck angle	0.913	0.867	0.942	Almost perfect
Glenoid offset	0.870	0.758	0.923	Almost perfect
Subscapularis angle	0.839	0.728	0.900	Almost perfect
Infraspinatus angle	0.904	0.858	0.935	Almost perfect
Resultant angle	0.786	0.626	0.869	Moderate
Humeral offset	0.913	0.839	0.949	Almost perfect
Glenoid vault tapering	0.813	0.724	0.873	Almost perfect
Protraction	0.871	0.792	0.920	Almost perfect
Scapular internal rotation	0.918	0.855	0.952	Almost perfect
Scapular upward rotation	0.943	0.908	0.965	Almost perfect
Scapular translation	0.620	0.390	0.764	Moderate
Scapular tilt	0.940	0.892	0.965	Almost perfect

ICC, intraclass correlation coefficient.

Comparison between patients with primary OA and control group

Means values of measurements and their statistical comparisons between case and control groups are demonstrated in Table III, Table IV, Table V.

Table III.

Means values of measurements and their statistical comparisons between type A glenoid and control group.

	Type A	Control A	P value
Number of shoulders (n)	24	24	-
Glenoid version according to the method of Friedman (°)	4.3 ± 4.1	5.0 ± 4.5	.482
Glenoid version relative to scapular blade axis (°)	3.5 ± 11.5	2.8 ± 5.1	.391
Glenoid inclination (°)	4.3 ± 5.7	6.7 ± 5.4	.205
Glenohumeral subluxation index	0.51 ± 0.03	0.51 ± 0.03	.661
Scapulohumeral subluxation index	0.56 ± 0.06	0.56 ± 0.05	.970
Neck angle (°)	171.3 ± 3.6	172.8 ± 4.2	.147
Glenoid offset (mm)	4.7 ± 2.0	3.9 ± 2.4	.175
Subscapularis angle (°)	74.2 ± 5.4	69.5 ± 5.1	.004
Infraspinatus angle (°)	101.0 ± 2.8	101.1 ± 2.9	.872
Resultant angle (°)	87.6 ± 3.4	85.5 ± 3.5	.020
Humeral offset (mm)	3.7 ± 4.1	2.5 ± 4.2	.320
Protraction (°)	93.7 ± 4.1	90.4 ± 4.8	.244
Scapular internal rotation (°)	42.7 ± 4.0	44.3 ± 5.0	.226
Scapular upward rotation (°)	10.9 ± 4.2	12.8 ± 5.3	.174
Scapular translation (°)	79.5 ± 5.5	81.8 ± 8.3	.333
Scapular tilt (°)	10.1 ± 6.8	17.8 ± 10.8	.044

Bold P values are statically significant (P < .05).

Table IV.

Mean values of measurements and their statistical comparisons between type B glenoid and control group.

	Type B	Control B	P value
Number of shoulders (n)	30	30	-
Glenoid version according to the method of Friedman (°)	13.1 ± 4.0	4.5 ± 4.2	<.001
Glenoid version relative to scapular blade axis (°)	12.2 ± 9.7	2.2 ± 5.1	<.001
Glenoid inclination (°)	2.8 ± 5.1	5.9 ± 3.9	.007
Glenohumeral subluxation index	0.55 ± 0.06	0.50 ± 0.03	<.001
Scapulohumeral subluxation index	0.68 ± 0.07	0.54 ± 0.04	<.001
Neck angle (°)	172.5 ± 5.4	171.8 ± 4.1	.571
Glenoid offset (mm)	4.1 ± 2.9	4.4 ± 2.2	.660
Subscapularis angle (°)	79.5 ± 7.3	68.9 ± 3.8	<.001
Infraspinatus angle (°)	97.0 ± 3.5	101.6 ± 3.4	<.001
Resultant angle (°)	88.2 ± 4.0	85.3 ± 3.4	.009
Humeral offset (mm)	−3.4 ± 5.6	3.4 ± 4.2	<.001
Protraction (°)	91.3 ± 2.8	91.1 ± 4.4	.963
Scapular internal rotation (°)	40.6 ± 5.6	44.7 ± 4.0	.020
Scapular upward rotation (°)	8.3 ± 3.7	14.5 ± 4.4	<.001
Scapular translation (°)	79.2 ± 7.0	79.3 ± 6.6	.964
Scapular tilt (°)	11.6 ± 5.6	15.7 ± 6.4	.032

Bold P values are statically significant (P < .05).

Table V.

Mean values of measurements and their statistical comparisons between type C glenoid and control group.

	Type C	Control C	P value
Number of shoulders (n)	5	5	-
Glenoid version according to the method of Friedman (°)	34.2 ± 9.5	3.9 ± 3.5	<.001
Glenoid version relative to scapular blade axis (°)	29.5 ± 9.5	1.3 ± 3.9	.002
Glenoid inclination (°)	−6.3 ± 5.4	6.7 ± 2.7	.016
Glenohumeral subluxation index	0.53 ± 0.13	0.51 ± 0.04	.703
Scapulohumeral subluxation index	0.86 ± 0.07	0.56 ± 0.05	<.001
Neck angle (°)	166.6 ± 5.3	171.1 ± 2.2	.236
Glenoid offset (mm)	7.5 ± 3.3	4.7 ± 1.2	.214
Subscapularis angle (°)	76.3	69.1	.236
Infraspinatus angle (°)	84.0 ± 9.5	101.6 ± 2.3	.017
Resultant angle (°)	80.1 ± 9.8	85.1 ± 2.7	.367
Humeral offset (mm)	−8.0 ± 4.0	3.5 ± 3.3	.015
Protraction (°)	93.4 ± 1.4	87.5 ± 2.8	.159
Scapular internal rotation (°)	47.7 ± 9.4	45.6 ± 7.3	.839
Scapular upward rotation (°)	4.8 ± 6.3	12.6 ± 2.6	.164
Scapular translation (°)	77.1 ± 5.7	64.5 ± 19.7	.421
Scapular tilt (°)	10.9 ± 10.2	29.7 ± 9.2	.235

Bold P values are statically significant (P < .05).

Type A glenoids showed no differences compared to their control group except significantly higher SSC angle, higher resultant RC vector, and lower scapular tilt.

Type B glenoids had higher glenoid version, lower glenoid inclination, higher glenohumeral and scapulohumeral subluxation indices, higher SSC angle, lower ISP angle, higher resultant RC vector, and more posterior humeral offset in comparison with their control group.

Type C glenoids showed significantly higher glenoid version, lower glenoid inclination, higher scapulohumeral subluxation index, posterior humeral offset, lower ISP angle with no significant change in the resultant RC vector in comparison with their control group.

Comparison according to Walch glenoid type

See mean values of measurements and their statistical comparisons among the Walch glenoid types in Table VI.

Table VI.

Means values of measurements and their statistical comparisons between the glenoid types according to Walch.

	OA groups			P value
	A	B	C	A/B	A/C	B/C
Number of shoulders (n)	24	30	5	-	-	-
Glenoid version according to the method of Friedman (°)	4.3 ± 4.1	13.1 ± 4.0	34.2 ± 9.5	<.001	<.001	<.001
Glenoid version relative to scapular blade axis (°)	3.5 ± 11.5	12.2 ± 9.7	29.5 ± 9.5	<.001	<.001	.046
Glenoid inclination (°)	4.3 ± 5.7	2.8 ± 5.1	−6.3 ± 5.4	.894	<.001	.003
Glenohumeral subluxation index	0.51 ± 0.03	0.55 ± 0.06	0.53 ± 0.13	.135	1.0	1.0
Scapulohumeral subluxation index	0.56 ± 0.06	0.68 ± 0.07	0.86 ± 0.07	<.001	<.001	<.001
Neck angle (°)	171.3 ± 3.6	172.5 ± 5.4	166.6 ± 5.3	1.0	.153	.041
Glenoid offset (mm)	4.7 ± 2.0	4.1 ± 2.9	7.5 ± 3.3	1.0	.124	.032
Subscapularis angle (°)	74.2 ± 5.4	79.5 ± 7.3	76.3 ± 9.6	.019	1.0	.984
Infraspinatus angle (°)	101.0 ± 2.8	97.0 ± 3.5	84.0 ± 9.5	.002	<.001	<.001
Resultant angle (°)	87.6 ± 3.4	88.2 ± 4.0	80.1 ± 9.8	1.0	.003	.001
Humeral offset (mm)	3.7 ± 4.1	−3.4 ± 5.6	−8.0 ± 4.0	<.001	<.001	.178
Protraction (°)	93.7 ± 4.1	91.3 ± 2.8	92.4 ± 1.4	.171	1.0	1.0
Scapular internal rotation (°)	42.7 ± 4.0	40.6 ± 5.6	47.7 ± 9.4	.825	.416	.119
Scapular upward rotation (°)	10.9 ± 4.2	8.3 ± 3.7	4.8 ± 6.3	.237	.077	.591
Scapular translation (°)	79.5 ± 5.5	79.1 ± 7.0	77.1 ± 5.7	1.0	1.0	1.0
Scapular tilt (°)	10.1 ± 6.8	11.6 ± 5.6	10.9 ± 10.2	.830	.334	.743

Bold P values are statically significant (P < .05).

OA, osteoarthritis.

The measurements of the glenoid version according to Friedman et al¹³ and relative to scapular blade axis showed both significantly higher retroversion of type B in comparison to A as well as type C in comparison to A and B.

Glenoid inclination was significantly lower in type C in comparison to A and B. However, there was no significant difference in glenoid inclination between type A and B.

The subluxation indices showed a significant higher scapulohumeral subluxation in type B in comparison to A, and in type C in comparison to A and B. However, there was no significant difference in glenohumeral subluxation indices between the 3 groups.

Evaluation of the neck angle and glenoid offset revealed a significant difference between types B and C, indicating a lower neck angle and higher anterior offset in type C glenoids.

Significant differences were also found in comparison of humeral offset: In types B and C, the midpoint of the humeral head was posterior to the scapular blade axis with a significant distance from type A.

The ISP angle showed significantly higher values in type A in comparison to types B and C, as well as in type B in comparison to type C. The SSC angle was significantly higher in type B than in type A glenoids. The resultant angle of pull in relation to the glenoid surface was significantly higher in types A and B compared to type C.

Evaluation of the scapulothoracic orientation did not demonstrate any significant differences between types A, B, and C.

Discussion

The association of eccentric OA of the shoulder with several osseous structural parameters including glenoid articular surface shape,^21,22 acromion shape,⁴ and humeral retrotorsion³⁹ as well as soft tissue characteristics including RC muscle volume² and atrophy¹¹ have been investigated. Although the goal was to identify intrinsic risk factors for the development of primary eccentric OA, the cause-effect relationship cannot be established easily. In addition, extrinsic factors such as heavy loading of the shoulder in weightlifters or combat sport athletes also have to be considered as risk factors.³⁸

The increased glenoid retroversion in eccentric OA seen in our data is supported by previously reported observations.²⁵ Different studies attempted to recenter the humeral head with correction of glenoid retroversion but could not achieve satisfying results considering the humeral subluxation—the head remained in the back most cases.^14,35 Although the glenoid vault model analysis to predict premorbid glenoid morphology has shown that glenoid version of the pathologic osteoarthritic shoulder did not differ significantly from healthy shoulders,⁴⁰ other authors reported increased retroversion in B2 glenoids compared with healthy controls.²² Currently, no consensus exists in the literature regarding the relationship between glenoid retroversion and humeral head subluxation and the causes for preosteoarthritic static posterior subluxation of the humeral head—it is likely multifactorial and related to both bony and soft tissue factors.²² Recently, Terrier et al found a strong correlation between glenoid version and scapulohumeral subluxation, showing an association of each degree of glenoid version with a percentage of subluxation in the same orientation.⁴⁷ However, it is important to underline that the glenohumeral subluxation index may not be the optimal reference in evaluating subluxation.¹⁰ The extent of glenohumeral subluxation is not closely related to the degree of glenoid retroversion.²⁵ The humeral head can indeed perfectly align with the glenoid fossa but may still be misaligned with the scapula. This was also confirmed by our findings.

In contrast to the previously noted bony abnormalities in patients with type C glenoids, such as increased anterior glenoid offset,^1,48 type B glenoids did not demonstrate a significant increase in anterior offset. This finding, based on glenoid vault tapering results, indicates a greater anterior offset in type C glenoids compared to types A, B, or controls. The commonly held belief that the humeral head shifts posteriorly in patients with eccentric OA appears to be confirmed, as type B glenoids showed a higher posterior humeral offset. These findings suggest that static posterior subluxation of the humeral head is likely multifactorial, with a predominant influence from soft tissue factors that may pull the humeral head posteriorly and contribute to posterior erosion.

The role of glenoid orientation in terms of inclination in the OA development remains controversial.^10,53 Both type B and C glenoids showed a significantly lower glenoid inclination in comparison with their control groups. As depicted in previous studies, type C patients had significantly more inferior inclination than type B3, recently reported by Chan et al,⁷ which may be attributed to dysplasia and deficient posteroinferior ossification.³⁷ Chalmers et al showed that severe OA is associated with a more inferiorly inclined glenoid resulting in more compressive joint reaction forces.⁶ However, other previous studies demonstrated no difference in glenoid inclination between normal patients and patients with OA.^40,42 Codsi et al indicated that the glenoid inclination was similar to the patients’ nonpathologic contralateral shoulder and glenoid inclination in the nonpathologic shoulder of these patients was not significantly different from those in normal control cadaver shoulders, supporting that patients with acquired posterior glenoid bone loss secondary to OA do not have altered glenoid inclination, which would predispose them to a pathologic state.^9,10,40

Currently, there is no evidence in literature regarding the RC action lines in the development of concentric or eccentric OA. Our study intended to identify significant differences in the vectors of the force couple, but a look at our results cannot reveal a clear trend. The role of the RC forces across the scapulohumeral joint is not fully understood. Literature shows evidence of a significantly higher fatty infiltration of the posterior RC with higher glenoid retroversion and posterior subluxation.^3,11,18,51 The authors also mentioned that we currently do not know if changes to the RC in humeral head subluxation are the cause or effect of the process. Werthel et al did not observe any correlation between the anteroposterior muscle volume ratio of the RC and glenoid parameters (retroversion, Walch classification, and humeral subluxation) in patients with OA.⁵² In our study, we also did not observe any clear asymmetries in RC action lines between case and control groups. So, the humeral head stays centered to the glenoid fossa due to increase in glenoid version caused by posterior erosion in type B glenoids even if the scapulohumeral measurement shows clear decentering (Fig. 3).

Figure 3.

Axial CT scan of a 61-year-old patient with primary osteoarthritis and Walch type B2 glenoid, where the humeral head shows clear posterior scapulohumeral decentering while the humeral head sits centered in the concavity of the neoglenoid. CT, computed tomography.

Interestingly, the 3D measurements of the scapulothoracic orientation could not obtain significant differences between the Walch types A, B, and C. For C1 glenoids, various studies identified less scapular upward rotation, greater scapular internal rotation, and reduced anterior tilt in patients with glenohumeral instability during elevation of the arm in the scapular plane.^24,34,44,46 However, there is a wide range of variations in study samples and methodologies, making it hard to evaluate them appropriately, and there are no studies in the literature describing such scapulothoracic orientation changes in relation to OA.

This study has some limitations. First, a static evaluation based on motionless shoulders imaged by CT was conducted, and most of the measurements were in 2D. Despite our standardized method to estimate each RC tendon’s anatomical insertion and related muscle force vector, magnetic resonance imaging and 3D evaluation could have improved our measurements. Second, only vectors were measured for the line of pull of the force couple, but other soft tissue components were not evaluated. As mentioned in the previously described measurement of scapulothoracic orientation by Moroder et al,³⁰ the 3D measurements in this study were obtained with the patients in the supine position, which shifts the scapula toward retraction. Therefore, protraction and scapular internal rotation might be lower in this study than we would have expected in a standing position. Third, the hand dominance of the patients is unknown, and it would have been interesting to determine whether hand dominance affected the measurements performed in the present study. Fourth, the study was conducted retrospectively, and measurements were not correlated with the clinical condition of the patients.

Conclusion

Glenoid type A shows minimal differences compared to a healthy control group. In contrast, shoulder joints with glenoid type B or C show significant differences of scapulohumeral centering in glenoid version, humeral offset and glenoid inclination compared to controls. Furthermore, despite the changes of SSP and ISP angles in type B glenoids, the humeral head stays centered to the glenoid fossa through compensation of multiple factors.

Disclaimers:

Funding: No funding was disclosed by the authors.

Conflicts of interest: The authors, their immediate families, and any research foundation with which they are affiliated have not received any financial payments or other benefits from any commercial entity related to the subject of this article.