Radiologic features of idiopathic pre-osteoarthritic posterior subluxation of the humeral head

Abstract

Background

Pre-osteoarthritic posterior subluxation of the humeral head (PPSHH), also referred to as a Walch B0 glenoid, is characterized by pathologic posterior subluxation of the humeral head (HH) without degenerative bony changes on plain radiographs. The purpose of this study was to describe the imaging findings of PPSHH and to assess the relationship between glenoid retroversion and HH subluxation.

Methods

This was a retrospective case series of patients with symptomatic PPSHH. Retroversion was measured on magnetic resonance imaging based on both Friedman’s line and the scapular axis line. HH subluxation was measured based on the glenohumeral index (GHI) and the scapulohumeral index (SHI). The relationship between retroversion and HH subluxation was evaluated. The difference in glenoid cartilage thickness between the anterior and posterior aspects of the glenoid was compared.

Results

Eight patients were included (mean age, 34.0±4.7 years). Mean retroversion was 17.2°±5.8° based on Friedman’s line and 14.0°±5.6° based on scapular axis line. Mean HH subluxation was 63.7%±6.2% based on GHI and 77.5%±9.9% based on SHI. Cartilage thickness was less in the posterior glenoid compared to anterior glenoid, and all patients had selective chondral wear on the posterior glenoid with a posterior labral tear. There was a direct relationship between the degree of glenoid retroversion and HH subluxation.

Conclusions

PPSHH is characterized by posterior glenoid chondral thinning and posterior labral tears in the setting of HH subluxation without erosive osseous changes. This pathology can present with varying degrees of retroversion and HH subluxation, and increasing retroversion is associated with more severe subluxation.

Level of evidence

IV.

INTRODUCTION

Primary glenohumeral osteoarthritis (GHOA) is characterized by progressive cartilage loss and bony erosion resulting in pain and limitation of function [1-4]. Although the specific cause of primary GHOA is not well understood, extensive research has been conducted exploring morphological variations of the glenohumeral joint as well as radiographic signs of GHOA progression [5-8]. Walch et al. [9] first described static posterior subluxation of the humeral head in the absence of posterior glenoid erosion and termed it B0 glenoid. In their series of 13 patients, presenting symptoms included stiffness, pain, locking of the joint, and negative provocative testing for instability. All patients demonstrated progression of osteoarthritis on follow-up [9]. This preliminary stage of primary GHOA has subsequently been termed pre-osteoarthritic posterior subluxation of the humeral head (PPSHH), or Walch B0 glenoid [7]. Importantly, PPSHH is characterized by static posterior subluxation (dynamic in the initial stages) in the absence of any of the degenerative or erosive osseous changes of the posterior glenoid seen with other types of B glenoid morphology. B0 glenoid varies from Walch B1 glenoid, which is characterized by posterior subluxation of the HH and degenerative changes in the posterior glenoid including subchondral sclerosis, osteophytes, or edema-like bone marrow changes [4,9].

The underlying etiology of primary GHOA is poorly understood. The description of PPSHH has furthered the knowledge of primary GHOA progression. Prior descriptions of the natural progression of osteoarthritis in the type B glenoid suggested that it proceeds sequentially from posterior wear to glenoid retroversion and HH subluxation [4,10]. Furthermore, it has been theorized that a dynamic, arm position-dependent subluxation may precede the static subluxation identifiable on axial computed tomography (CT) [7,11]. It is unclear whether PPSHH is secondary to isolated posterior chondral loss in the setting of excessive glenoid retroversion with increased posteriorly directed forces, or transverse plane rotator cuff imbalances, or a combination of bony and soft tissue abnormalities [5,7,12-14]. The goal of early identification of this pathologic process is to determine whether treatment options like glenoid osteotomy, anterior capsular release or posterior bone block can halt or delay progression to more severe GHOA [6,7]. The finding of PPSHH or B0 glenoid have been described on plain radiographs and CT but not on magnetic resonance imaging (MRI), where cartilage is seen much better. Therefore, the purpose of this study was to describe the MRI findings associated with PPSHH and to assess the relationship between glenoid retroversion and the extent of HH subluxation.

METHODS

The study protocol was reviewed and approved by the Institutional Review Board of New York University Langone Health (No. i06-580). The requirement for informed consent was waived by the Board.

This was a retrospective case series consisting of patients treated by a single fellowship-trained shoulder and elbow surgeon (MV). The inclusion criterion was a diagnosis of PPSHH, which was defined as posterior subluxation of HH without any posterior glenoid erosion. Exclusion criteria consisted of glenoids with posterior glenoid bony changes (subchondral sclerosis, osteophytes-Walch B1), traumatic posterior shoulder instability, any non-PPSHH diagnosis as the cause of the patient’s symptoms, or history of connective tissue disorder.

A total of eight patients were included in this cohort. Basic demographic information was recorded. All patients underwent radiographs and MRI of the affected shoulder using standard institutional protocols. Plain radiographs and MRIs were reviewed independently by two of the authors (MTK and AB). X-ray evaluation was performed to ensure the lack of radiographic signs of osteoarthritis based on standard anterior-posterior, scapular-Y, and axillary views). MRIs were obtained without contrast enhancement using a minimum 1.5T scanner with the arm in a neutral position.

The degree of glenoid version and HH subluxation were each measured using two different previously established methods [4,10,15,16]. Glenoid version was measured relative to both Friedman’s line (FL) and the scapular axis line. The same axial slice of the imaging study was used for both measurements. As originally described by Friedman et al. [10], glenoid version can be measured relative to a line drawn from the medial tip of the scapula to the anterior-to-posterior midpoint of the glenoid (i.e., FL). A second line is drawn tangent to the anterior and posterior margins of the glenoid rim (not including osteophytes). Glenoid version is represented by the angle between the line tangent to the glenoid rim and the perpendicular of FL (Fig. 1A). Alternatively, glenoid version may be measured relative to the scapular axis, which is a line drawn from the medial tip of the scapula to the most medial aspect of the glenoid vault then extended through the glenoid surface. The angle between the perpendicular of the scapular axis line and a line tangent to the glenoid rim represents the glenoid version (Fig. 1B) [15].

Fig. 1.

Glenoid retroversion as measured based on Friedman’s line (A) and the scapular axis line (B).

Glenohumeral subluxation was quantified based on both the glenohumeral index (GHI) and the scapulohumeral index (SHI). The same axial slice was used for both measurements of glenohumeral subluxation. To obtain the GHI, a best fit circle is first drawn around the humeral head, aligned with the articular surface. A line tangent to the anterior and posterior margins of the glenoid rim is drawn, and then this line is bisected with a perpendicular line starting at the midpoint of the glenoid and extending laterally through the humeral head. The distance (measured perpendicular to the glenoid bisecting line) from the bisecting line to the posterior edge of the best fit circle is divided by the diameter of the HH best fit circle to obtain the GHI (Fig. 2A). An index greater than 55% indicates posterior subluxation [4]. Alternatively, glenohumeral subluxation can be quantified using the SHI, in which the distance (measured perpendicular to FL) from FL to the posterior edge of the HH best fit circle divided by the diameter of the best fit circle gives the SHI (Fig. 2B) [16].

Fig. 2.

Humeral head subluxation as measured by the glenohumeral index (A) and the scapulohumeral index (B). The longer length red line represents the cross-sectional diameter of the humeral head as measured using the best fit circle method. The shorter length red line represents the section of the humeral head diameter posterior to the glenoid bisecting line (A) or posterior to the Freidman’s axis (B).

For each patient, glenoid cartilage thickness was measured on MRI [17]. Two coronal slices were chosen, one at the junction of the anterior third and middle third of the glenoid and another at the junction of the middle third and posterior third of the glenoid. On each slice, cartilage thickness was measured superiorly, centrally, and inferiorly, and these three values were averaged to determine the mean cartilage thickness at the anterior and posterior glenoid.

Statistical Analysis

Inter-observer reliability

Two independent raters measured glenoid retroversion using FL and Scapular Axis-Line (SAL) methods, and HH subluxation using the GHI and SHI. For each parameter, inter-observer reliability was assessed using the intraclass correlation coefficient (ICC), which was estimated with a two-way random effects model for absolute agreement of single measures (FL, SAL, GHI, and SHI). Higher ICC values indicate stronger agreement between the two raters.

Association between retroversion and subluxation

To evaluate the association between glenoid retroversion and HH subluxation, we averaged values across the two raters for each patient. Because of the small sample size, we used Spearman’s rank correlation to assess associations of FL retroversion with both GHI and SHI, and of SAL retroversion with both GHI and SHI.

Glenoid cartilage thickness

Anterior and posterior glenoid cartilage thickness were compared within patients using a paired Wilcoxon signed-rank test. Effect size was quantified using the rank-biserial correlation. For descriptive purposes, we report medians and interquartile ranges (IQRs) for anterior and posterior thickness, and for their paired differences. All analyses were two-tailed and set at the P<0.05 threshold. Analyses were conducted using R Studio for Windows version 4.4.2 (R Foundation for Statistical Computing).

RESULTS

Demographics

Seven of the eight patients who met inclusion criteria and were included in this study were male, and the mean age at the time of presentation was 34.0±4.7 years. The patients in this cohort were generally young and active, and initially presented with chronic or acute-on-chronic shoulder pain that was made worse with overhead throwing or lifting movements. Patients reported clicking and difficult with overhead lifting.

Radiographic Findings

Diagnostic imaging consistently showed posterior subluxation of the HH with thinning of the posterior glenoid cartilage and posterior labral tearing, but without erosive or degenerative osseous changes. The mean retroversion was 17.2°±5.8° (range, 8.6°–24.8°) based on the FL method and 14.0°±5.6° (range, 6.9°–22.8°) based on the scapular axis line method. The mean posterior HH subluxation was 63.7%±6.2% (range, 56.1%–74.2%) based on the GHI method and 77.5%±9.9% (range, 65.8%–91.0%) based on the SHI method (Fig. 3).

Fig. 3.

Glenoid retroversion (A) and humeral head subluxation (B) in patients with pre-osteoarthritic posterior subluxation of the humeral head. FL: Friedman’s line, SAL: scapular axis line, SHI: scapulohumeral index, GHI: glenohumeral index.

Inter-observer Reliability

Inter-observer agreement was good to excellent for retroversion and subluxation measures. ICCs (95% CIs) were 0.79 (0.24–0.96) for FL retroversion, 0.79 (0.22–0.95) for SAL retroversion, 0.67 (–0.03 to 0.93) for GHI, and 0.82 (–0.06 to 0.97) for SHI.

Association between Retroversion and Subluxation

Spearman’s correlations demonstrated strong positive associations between retroversion and HH subluxation (Table 1, Fig. 4).

Table 1.

Spearman correlations between retroversion and humeral head subluxation

Comparison	n	ρ (rho)	P-value
FL vs. GHI	8	0.79	0.021
FL vs. SHI	8	0.95	<0.001
SAL vs. GHI	8	0.83	0.010
SAL vs. SHI	8	0.93	<0.001

Spearman’s rank correlation coefficients (ρ) between retroversion measurements (FL and SAL) and humeral head subluxation indices (GHI and SHI). Correlations were calculated using rater-averaged values per patient.

FL: Friedman’s line, GHI: glenohumeral index, SHI: scapulohumeral index, SAL: scapular axis line.

Fig. 4.

Series of simple linear regression models demonstrating the relationship between glenoid retroversion and posterior humeral head subluxation in patients with pre-osteoarthritic posterior subluxation of the humeral head based on different methods of measurement for each parameter. There was a strong positive association between degree of retroversion and subluxation irrespective of measurement methodology. (A) Friedman’s line (FL) vs. glenohumeral index (GHI). (B) FL vs. scapulohumeral index (SHI). (C) Scapular axis line (SAL) vs. GHI. (D) SAL vs. SHI.

Glenoid Cartilage Thickness

Anterior glenoid cartilage thickness (median [IQR], 2.05 mm [1.98–2.13]) was significantly greater than posterior thickness (1.25 mm [1.10–1.45]); the median paired difference was 0.60 mm (0.45–0.90) (Table 2). The Wilcoxon signed-rank test confirmed this difference (W=36, P=0.014), with a large effect size (rank-biserial r=0.89).

Table 2.

Anterior versus posterior glenoid cartilage thickness (n=8)

Anterior (mm)	Posterior (mm)	Difference (mm)	W statistic	P-value	Effect size (r)
2.05 (1.98–2.13)	1.25 (1.10–1.45)	0.60 (0.45–0.90)	36	0.014	0.89

Comparison of anterior and posterior glenoid cartilage thickness. Values are reported as median (interquartile range). The Wilcoxon signed-rank test was used for paired comparisons, with effect size expressed as rank-biserial correlation (r).

Clinical and Intraoperative Findings

Three of the eight included patients underwent arthroscopic shoulder debridement and posterior labral stabilization as attempted first-line treatment based on patient preference. In this cohort, soft tissue-only procedures failed consistently, resulting in recurrence of pain and limited shoulder range of motion. One of the patients who underwent initial treatment with arthroscopic labral repair (at an outside hospital) subsequently underwent posterior bone block augmentation following the return of symptoms (patient 6) (Table 3). This patient’s symptoms did not improve following posterior labral repair with an appropriate course of postoperative physical therapy, and continued to worsen with activities of daily living. Revision surgery with a posterior bone block was thus recommended, which resulted in resolution of his pain and symptoms.

Table 3.

Details of patients presenting with PPSHH

Patient	Age (yr)	Sex	Surgery	Laterality	Retroversion (FL, °)	Retroversion (SAL, °)	Subluxation (GHI, %)	Subluxation (SHI, %)
1	36	Male	None	Left	8.6	6.9	61.7	69.1
2	34	Male	None	Right	13.3	12.8	57.4	69.2
3	34	Male	Posterior bone block	Left	13.0	7.8	56.0	65.8
4	34	Male	Posterior bone block	Left	24.4	22.8	74.2	90.0
5	27	Male	Posterior bone block	Right	21.3	13.9	66.2	84.9
6	28	Male	Posterior labral repair followed by posterior bone block	Right	24.8	20.9	69.8	91.0
7	38	Male	Posterior labral repair	Left	15.4	14.2	64.3	75.9
8	41	Female	Posterior labral repair	Left	17.0	13.0	59.8	74.2

Retroversion values represent degrees of glenoid retroversion. Subluxation values represent the percentage of humeral head posterior subluxation.

PPSHH: pre-osteoarthritic posterior subluxation of the humeral head, FL: Friedman’s line, SAL: scapular axis line, GHI: glenohumeral index, SHI: scapulohumeral index.

Arthroscopic imaging for the patients in this cohort demonstrated grade 2-4 chondral loss in the posteroinferior quadrant of the] glenoid with degenerative posterior labral tear. The HH cartilage was largely preserved and anterior glenoid did not have significant glenoid wear. Four patients underwent posterior bone block augmentation using distal tibia allograft, which increased the glenoid arc and provided support to the subluxated HH and resulted in good outcomes (Table 3).

DISCUSSION

PPSHH is a distinct entity characterized by posterior static subluxation without any erosive bony changes in the glenoid. In this study, we present the radiographic findings of PPSHH on MRI in a small cohort of eight patients. All patients with PPSHH demonstrated posterior glenoid cartilage thinning, posterior HH subluxation and posterior labral tear on MRI. Intraoperative findings (shoulder arthroscopy) demonstrated cartilage loss in the posterior glenoid with posterior labral tear and preservation of cartilage in the anterior half of the glenoid. Furthermore, we demonstrate that PPSHH can present with varying degrees of glenoid retroversion and varying amounts of HH subluxation, and that increasing glenoid retroversion generally tends to be associated with more severe subluxation.

The mean glenoid retroversion was 17.2° based on the FL method and 14.0° based on the scapular axis line method. Mean posterior HH subluxation was 63.7% based on the GHI method and 77.5% based on the SHI method. While there is a paucity of existing literature describing the typical glenohumeral morphology of patients with PPSHH, our radiographic findings are consistent with prior reports. Walch et al. [9] first described PPSHH in a case series of 13 patients with mean glenoid retroversion of 14.5° (range, 6°–27°) measured via FL method and mean GHI subluxation of 65% (range, 58%–75%). Similarly, Akgün et al. [5] reported radiographic findings of 14 shoulders in 12 patients with PPSHH. Mean glenoid retroversion based on the FL and scapular axis line were 19°±10° and 14°±10°, respectively. Mean subluxation based on GHI and SHI were 58%±10% and 76%±10%, respectively. The authors also reported increased anterior offset of the glenoid vault and decreased glenoid neck angles in the PPSHH group.

Currently there is conflicting evidence regarding the association between glenoid retroversion, HH subluxation, and degenerative changes in PPSHH. In the initial description, Walch et al. [9] identified increased glenoid retroversion as the direct cause of static posterior HH subluxation prior to the onset of glenoid erosion on radiographs or CT. This was in contrast to the underlying pathophysiologic process described by Neer et al. [3], in which posterior glenoid erosion preceded HH subluxation. Our findings support the hypothesized sequence of pathologic events proposed by Walch. The patients in our cohort demonstrated glenoid retroversion with associated posterior subluxation and selective posterior cartilage thinning, but without bony erosive changes. However, not all shoulders go through the same PPSHH/B0 to B1 to B2 and B3 sequence.

Gerber et al. [12] found no association between increased glenoid retroversion and articular subluxation as measured by the FL method and the glenohumeral index, respectively. Hoenecke et al. [13] failed to demonstrate a correlation between retroversion measured via the scapular axis line and the GHI. Yet other studies have demonstrated a strong relationship between glenoid retroversion and the scapulohumeral index. Terrier et al. [14] reported a high correlation between retroversion and subluxation and found that each degree of retroversion induced one percentage point of HH posterior subluxation. Akgün et al. [5] also demonstrated a significant relationship between posterior subluxation as measured via the SHI and excessive glenoid retroversion. In contrast, no association between the GHI and glenoid retroversion was found. While rotator cuff asymmetry, genetic factors, and other bony abnormalities have been posited to play a role in the pathogenesis of PPSHH [5,18,19], our data supports the assertion that increasing glenoid retroversion may indeed be associated with more severe posterior subluxation and play an important role in the pathogenesis of PPSHH. Furthermore, it is unclear whether selective posterior chondral wear is a result of genetic predisposition or is secondary to increased biomechanical loading of posterior glenoid cartilage in patients with higher degrees of retroversion and posterior subluxation, resulting in initiation of arthritic changes at an early age.

This study also highlights the need for consensus on and standardized reporting of glenoid version and subluxation measurements. The scapular axis line theoretically provides a more accurate representation of the action plane of the rotator cuff compared to FL and may therefore represent a more clinically relevant description of glenoid retroversion. While average retroversion measurements differ based on the method used, both have demonstrated excellent intra-observer and inter-observer reliability [15]. With respect to HH subluxation, the SHI method generally demonstrates higher subluxation percentages compared to the GHI method [1,16]. It has been suggested that the SHI represents a more accurate measurement of articular subluxation, as it measures alignment in the plane of the scapula and rotator cuff. This is particularly true when significant glenoid erosion or dysplasia is present, as the HH may be posterior to the scapular axis but remain centered within the glenoid fossa [7,16]. However, as with glenoid retroversion, no consensus exists regarding the optimal measurement method for subluxation. Our data suggest that for this patient population, the relationship between glenoid version and HH subluxation is best described by retroversion measured using FL and subluxation measured using the SHI (which is measured relative to FL). However, the observed relationship was maintained regardless of the measurement technique used.

There is limited evidence on treatment options for PPSHH. In the present study, three of the eight patients underwent arthroscopic shoulder debridement and posterior labral stabilization, all of which failed, resulting in recurrent pain and limited shoulder range of motion. Four patients, including one patient with failed labral repair, underwent posterior bone block augmentation using a distal tibia allograft to support the subluxated humeral head. Walch et al. [9] reported two patients who underwent posterior opening wedge osteotomy of the glenoid neck and posterior capsular imbrication. At a mean follow-up of 46 months, they noted worse constant scores, progressive osteoarthritic changes, and higher subluxation indexes. The authors [9] suggested that erosion of the posterior glenoid cartilage may have been severe enough at presentation to facilitate osteoarthritic progression despite intervention. In addition, they described two patients who underwent anterior arthrolysis with posterior capsulorrhaphy and one patient who underwent posterior glenoid augmentation with iliac crest autograft. All three patients demonstrated improvement in constant scores on follow-up [9]. Ortmaier et al. [20] reported 10 cases treated with posterior closing wedge osteotomy at a mean follow-up of 33 months. They noted improvement in constant scores in all shoulders and an increase in mean forward flexion from 117° preoperatively to 143°. As awareness and diagnosis of PPSHH increases, future research should assess whether bony and/or soft tissue procedures are successful in slowing the progression of GHA.

This study has certain limitations. First, the sample size of the cohort was small and MRI and clinical findings will require validation in a larger cohort. Prior studies have reported their findings in small cohorts, suggesting that PPSHH is not a common phenomenon in GHOA. Additionally, due to the small sample size, the statistical significance of reported findings carries a risk of bias. Second, although we have described intraoperative findings in this study, these findings are not available for all patients in the cohort. Third, we could not provide patient-reported outcome measures or long-term clinical outcomes in these patients. Lastly, we are unable to evaluate the radiographic progression of PPSHH over time due to a lack of longitudinal imaging data.

CONCLUSIONS

Pre-osteoarthritic posterior subluxation of the humeral head is a distinct entity. This study illustrates MRI findings characteristic of PPSHH namely posterior glenoid chondral thinning and posterior labral tears, without significant erosive osseous changes. Glenoid retroversion and HH subluxation demonstrate a linear relationship in PPSHH with increasing glenoid retroversion associated with more severe HH subluxation.