Introduction

The bursal-sided partial tear of the supraspinatus (SSP) tendon is a rare entity that has not been described in detail in the literature and its pathomechanism is still the subject of debate. A multifactorial etiology with a combination of extrinsic and intrinsic pathomechanisms is assumed.

Neer et al. suggested early on a connection between subacromial abrasion of the SSP tendon, its wear and tear, and subsequent rupture and thus they propagated the theory of the extrinsic pathomechanisms of rotator cuff tears (RCTs; [1]). This pathogenesis, initially described by Neer, is due to the so-called mechanical outlet or subacromial impingement (MOI). The MOI, also known as the “classic impingement syndrome,” is described as a mechanical narrowing of the subacromial space by extratendinous elements [2]. This structural constriction can be caused by anterolateral acromial spurs, osteophytes under the acromioclavicular joint (ACJ), or posttraumatic malpositions of the greater tubercle [1]. Neer et al. postulated that 100% of impingement syndromes and 95% of rotator cuff pathologies are the result of an MOI [1].

Codman suggested that RCTs are the consequence of degenerative intratendinous (intrinsic) changes [3]. These intrinsic mechanisms describe degenerative processes within the tendon itself that develop over time due to wear and tear, stress, or trauma. Genetic and pathological studies provide evidence that intrinsic factors play a role, but little is known about the specific extent of the involvement of these factors in RCTs [4,5,6].

It is assumed that full-thickness RCTs result from the progression of partial ruptures of the rotator cuff [6]. This hypothesis was also supported by studies using biomechanical models of impingement [7]. Yamanaka et al. showed that after 1 year, 28% of partial ruptures progressed to full-thickness ruptures and 80% increased in size [8]. Other authors also described progression, but to a lesser extent [9, 10]. However, the progression of partial ruptures to full-thickness ruptures seems to depend on the degree of rupture [9].

Subacromial impingement of the SSP tendon due to MOI and its involvement in the pathogenesis of SSP tears in general was a frequently reported occurrence in the past. However, the literature has been conflicting regarding the strength of the association between radiological parameters in the pathogenesis of SSP tears. Furthermore, to date there is only limited evidence supporting the association between bursal-sided partial tears of the SSP tendon and radiological parameters of MOI. The aim of this study was therefore to expand the previous scientific evidence on the bursal-sided partial tear of the SSP tendon in connection with the MOI.

Materials and methods

The acquisition of patient data was carried out from October 2014 to November 2020 at the ATOS Clinic Heidelberg and ATOS Clinic Munich. In total, 43 patients with a bursal-sided partial tear of the SSP tendon (group 1) were matched with 43 patients with a complete, isolated tear of the SSP tendon (group 2). Matching criteria were age and sex. The group allocation was based on rupture type (Table 1), which was classified intraoperatively.

Table 1 Inclusion and exclusion criteria for groups 1 and 2
Full size table

The standardized preoperative radiographs (anteroposterior and outlet view) were compared with regard to the following radiological impingement parameters (Fig. 1):

  • Critical shoulder angle (CSA)

  • Lateral acromial angle (LAA)

  • Acromial index (AI)

  • Acromiohumeral distance (AHD)

  • Acromion type according to Bigliani (ATB)

Fig. 1
figure 1

Standardized anteroposterior radiographs showing radiological impingement parameters: critical shoulder angle (a), lateral acromial angle (b), acromiohumeral distance (c), acromiohumeral index (d). GA glenoacromial distance, GH glenohumeral distance

Full size image

Radiographs were evaluated by a blinded examiner. To assess intrarater reliability, the same radiographs were reassessed by the same blinded person 3 months after the initial evaluation. To assess interrater reliability, the radiographs were analyzed by another blinded, independent examiner.

The CSA, LAA, and AHI were measured as described by Moor et al. [11], Banas et al. [12], and Nyffeler et al. [13], respectively, in true anteroposterior radiographs. The AHD was defined as the perpendicular distance from a line passing through the inferior border of the acromion to the most cranial point of the humeral head. Acromial shapes were assigned using the Bigliani et al. classification on outlet-view radiographs [14].

Statistical analysis

All statistical analyses were performed using SPSS version 29.0 (SPSS Inc., Chicago, IL, USA). First, all data were tested for normality using the Kolmogorov–Smirnov test. Descriptive analysis was used to evaluate various metric and nominal variables, as well as to calculate means and standard deviations. The comparison of categorical variables (acromion type) was carried out using the chi-square test. Continuous variables (CSA, LAA, AHD, AHI) were compared using the Mann–Whitney U test. Inter- and intrarater reliability was determined for continuous variables (CSA, LAA, AHD, AHI) using the intraclass correlation coefficient as well as using Cohen’s kappa for categorical variables (acromion type). Statistical significance was defined as p ≤ 0.05.

Results

Our study collective was made up of 70% male (n = 60) and 30% female (n = 26) patients. The average age of the patients at the time of surgery was 57.8 years (SD ± 9). Means of radiological impingement parameters for groups 1 and 2 with standard deviations and 95% confidence intervals are presented in Table 2. No statistically significant differences were found between groups 1 and 2 when comparing CSA, LAA, AHD, and AHI (Table 2). The distribution of the acromion types according to Bigliani is shown in Table 3. The same order of acromion type distribution was observed in group 1 and in group 2, but with a slight nonstatistically significant difference in frequency (p = 0.443).

Table 2 Descriptive statistics for groups 1 and 2
Full size table
Table 3 Acromial types according to Bigliani in groups 1 and 2
Full size table

The intra- and interrater reliability for CSA, LAA, AHD, AHI, and ATB were excellent (p < 0.001).

Discussion

In this study, the type II acromion was the most common acromion type in both groups. These results are consistent with the findings of most other studies on this subject [15,16,17]. Few studies have investigated the acromion type according to Bigliani (ATB) in partial tears. Kaur et al. found no relationship between tear type and acromion type, although they did not differentiate between bursal- and articular-sided partial lesions [18]. Only our study and the study by Pandey et al. provide information on Bigliani’s acromial type in bursal-sided partial tears of the SSP tendon [19]. Our results showed no significant difference in the frequency distribution of the ATB between group 1 and group 2. Pandey et al. also found no relationship between the ATB and tear type. Our study, as well as the work of Pandey et al., suggests that the Bigliani classification is not a parameter suitable for differentiating between the two tear types.

Since the introduction of the critical shoulder angle (CSA) in 2013, a frequent correlation has been found between an increased angle and full-thickness RCTs [11, 20,21,22,23,24,25,26,27]. Cunningham et al. showed that a large CSA was associated with full-thickness RCTs but not with partial tears [23]. Seo et al. showed in a retrospective analysis that a high CSA was more associated with full-thickness tears and articular-sided partial tears than with bursal-sided partial tears. They found a significant difference in CSA between the bursal group with 31.5° ± 4.6° compared to 34.7° ± 4.4° in the full-thickness tear group (p = 0.001; [28]). Pandey et al. found no significant difference in CSA between bursal and full-thickness RCTs [19]. Our results also showed no statistically significant difference in CSA between group 1 and group 2. The CSA of our patients was on average slightly above the cut-off value initially defined by Moor et al. Lin et al. described a CSA of 36° on average in 300 patients without RCTs. The CSA in both of our groups, as well as that in the study by Seo et al., is on average below the values described in other studies on RCTs and CSA [19, 26, 29]. This suggests that CSA is not associated with RCTs to the extent reported by other authors.

The correlation between a narrowing of the acromiohumeral distance, measured through AHD, and RCTs has been extensively described in the literature [30,31,32]. The cause of the narrowing is a traction movement of the humeral head with displacement in the glenoid, due to cranial pull by the deltoid muscle in combination with failure of the superior rotator cuff [32, 33]. Traditionally AHD has been determined primarily by X‑rays [34, 35]. However, the radiographically measured AHD is not a reliable parameter for the true acromiohumeral distance, according to a review [36]. Our results showed very good inter- and intrarater reliability. There is little information in the literature about partial tears and AHD. In a study by Kaur et al., patients with full-thickness and partial-thickness tears had a smaller AHD than the control group without tears [18]. Both groups in our study had a similar AHD with values that were higher than those previously described in tear groups and are similar to those in healthy patients [18, 30, 35]. To our knowledge, our results provide the first data on AHD in bursal-sided partial tears of the SSP tendon.

The LAA was introduced by Banas et al. in 1995, who identified a correlation between a small LAA and the risk of impingement syndrome [12]. Since then, this association has been reported by other authors as well [16, 26, 37]. In their initial study of the LAA, Banas et al. described an angle of > 80° in asymptomatic patients, with patients with impingement syndrome symptoms showing a lower angle. In this study, patients with complete and partial ruptures were examined, and all patients with an LAA < 70° had complete RCTs [12]. In a few studies, RCTs were associated with an angle of < 70°, leading some authors to present the LAA as an independent factor for RCTs [12, 16]. Our study showed no difference in the LAA between patients with partial and complete lesions. Furthermore, both groups had an average LAA of about 80° despite the presence of an RCT. Therefore, our results did not support the LAA as a factor for RCTs.

Nyffeler et al. were the first to describe the association between RCTs and an increased acromiohumeral index (AHI; [13]). Although this initial observation was made on magnetic resonance images, it has since been shown that the AHI can also be applied to X‑rays [12, 16, 38]. This association has been confirmed multiple times in the literature [39,40,41,42]. In these studies, the average AHI in patients with RCTs was 0.73, and 0.67 in the control group with an intact rotator cuff. In their comparison of patients with complete RCTs with healthy patients, Moor et al. established an AHI cut-off value of 0.68 [26]. Torrens et al. also showed an association between a large AHI and RCT in a study of 148 patients [41]. However, there is also literature that argues against the association between AHI and RCTs. Several authors found no correlation between AHI and RCT [40, 43, 44]. Lin et al. reported an increased AHI of an average of 0.77 in patients with intact rotator cuffs [44]. Other authors have examined the difference in AHI between different types of RCT. Kim et al. found a larger AHI in patients with complete RCTs than in those with articular-sided RCTs [39]. In 2016, Balke et al. described a difference in AHI between patients with degenerative and traumatic RCTs. Patients with degenerative tears had a significantly greater acromial roofing than those with traumatic tears [45]. To our knowledge, there is only one study to date that has compared AHI between complete and bursal-sided partial tears. Pandey et al. found no statistically significant difference in AHI between complete and bursal-sided partial tears [19]. Our results are in line with those of Pandey et al., as no significant difference in AHI was found between group 1 and group 2.

Interestingly, the primary hypothesis could not be confirmed. Our study showed that CSA, LAA, AHD, and AHI are not suitable parameters for differentiating between bursal-sided partial lesions and complete ruptures of the SSP tendon. This suggests a relationship between the two rupture types and supports the hypothesis of bursal-sided partial lesions as a precursor to complete supraspinatus tendon rupture. In addition, our results suggest that radiological impingement parameters are not a useful diagnostic tool for differentiating these two types of RCTs. We showed that the impingement parameters CSA, LAA, and AHD, which are frequently described in the literature, do not differ significantly from the healthy normal values described in the literature. This suggests that contrary to the often-described assumption in the literature in the past, acromion morphology and MOI are not the only primarily causative factors for supraspinatus tendon ruptures. Our results are supported by the clinical evidence of some authors in which anterior subacromial decompression did not lead to symptom relief and did not reduce the risk of re-rupture or rupture progression [46,47,48,49,50,51,52,53,54,55,56,57,58,59]. Involvement of MOI in the pathogenesis cannot be excluded, but our results support the assumption of a multifactorial pathogenesis of RCTs, with intrinsic and extrinsic factors, and reduce the importance of the influence of MOI.

Limitations

Routine radiographs were utilized in the study. Despite standardization, factors such as the positioning of the patient, the X‑ray tube, and the plate can influence the reliability of the images. Our study did not differentiate between degenerative and traumatic rotator cuff ruptures. Further, in this work, we did not distinguish between chronic, acute, or acute-on-chronic supraspinatus tendon ruptures.

Practical conclusion

  • Our study revealed no significant difference in the radiological impingement signs—critical shoulder angle, lateral acromial angle, acromiohumeral distance, acromiohumeral index, and acromion type according to Bigliani—between patients with partial tears on the bursal side and those with complete tears of the supraspinatus tendon.

  • We thus conclude that scapular anatomy alone, as assessed by radiological impingement signs, is insufficient to distinguish between the two types of ruptures.

  • Our results also suggest a reduced influence of mechanical outlet impingement on the pathogenesis of rotator cuff tears (RCTs), and thus offer supporting evidence of a multifactorial pathogenesis of RCTs involving extrinsic and intrinsic mechanisms.

  • Our findings are relevant for patient communication but they also add to a more concise understanding of factors predisposing to success/failure of rotator cuff repair in the community of sports medicine physicians and healthcare providers.