Methodological aspects of the acromiohumeral distance measurement with ultrasonography—Reliability and effects of extrinsic and intrinsic factors in overhead and non‐overhead athletes

To improve comparability and interpretation of acromiohumeral distance (AHD) measurements, consequences of varying methodological approaches and population specific effects have to be known. This study aimed to investigate the intra‐ and inter‐rater reliability of different AHD ultrasound image analysis approaches in asymptomatic overhead‐ and non‐overhead athletes. Furthermore, the impact of shoulder muscle activity as well as relationships between AHD and individual factors were examined in different measurement positions.


| INTRODUCTION
Overhead sports such as handball and volleyball are characterized by repetitive high-velocity overhead motion, imposing high stresses on tissues of the upper extremity. 1,2 In addition to the increased risk of traumatic injuries, 3,4 chronic overuse may lead to a high prevalence of pathologic shoulder conditions in overhead athletes (OHA). 4,5 Even the shoulders of asymptomatic OHA show structural abnormalities traced back to sport-specific loading. 5 Hence, early and simple identification of structural abnormalities by regular screening of such athlete's shoulders may be a valuable option to assist the prevention, diagnosis and rehabilitation of shoulder injuries.
The subacromial space, which has been shown to reduce during shoulder abduction and elevation, 6 plays an important role in the shoulder health of OHA whose sport-specific motions are frequently performed in shoulder abduction and external rotation. 7 Reduced subacromial space has been associated with subacromial impingement syndrome which may lead to structural damage of structures such as the subacromial bursa or the rotator cuff tendons. 8 To prevent an impingement of the rotator cuff tendons, it appears to be crucial for athletes to conserve the subacromial space as much as possible 9 and appropriate interventions may assist in this process.
A method to quantify the subacromial space by measuring the linear distance between the acromion and the humeral head, called the acromiohumeral distance (AHD), has been established using various radiological methods. 10 The advantages of ultrasonography (US) such as low costs, high practicability and the absence of radiation outdo other methods especially when applied in healthy participants. 10,11 To provide realistic data with AHD measurements, it is important to utilize reliable and valid measurement techniques. Several studies investigated the reliability of the US image acquisition and demonstrated high intra-rater and inter-rater reliability for this technique. [11][12][13] However, precise information on the subsequent image analysis process, including localizing the anatomical landmarks between the acromion and the humeral head on the US image, is missing in some studies. 10 In addition, studies differ in terms of AHD definition. For example, some studies measure the tangential or nearest distance from the humeral head to the tip of the acromion whereas others measure the point of entry of the tendon into the acoustic shadow to the humeral head. 14 However, reliability and accuracy of the image analysis may depend on the choice of specific landmarks, which has not been investigated sufficiently.
For a precise and reliable determination of the AHD factors that may affect the measurement should be taken into account. The AHD has been shown to change with muscle contraction. 15 Measurements of the AHD while the shoulder muscles are contracting when holding the weight of the arm or even an additional load may thus differ from measurements during which the arm is resting passively, and muscles are not actively contracting. However, the effect of the activity level of the shoulder muscles on the AHD measurement has not been described conclusively.
Repetitive loading commonly leads to sports-related structural adaptations. 16 In overhead athletes the shoulder range of motion (ROM) is one component that has been shown to adapt according to prolonged and repeated overhead motion. Adaptions primarily present as glenohumeral internal rotation deficit (GIRD) of the dominant shoulder, defined as a loss of internal rotation (IR) of at least 20 or greater compared to the contralateral shoulder. 17 Furthermore, external rotation (ER) weakness and lower ratios of ER to IR strength (ER/IR ratio) have been reported in OHA. These sports-related adaptations of the shoulder as a consequence of repetitive loading in OHA have been shown to be risk factors for overuse injuries. 16 Adaptive changes and their effects on AHD have not been investigated in experienced OHA on competitive amateur level. Those are in comparison to professional elite athletes a much larger but far less well investigated study population, being exposed to many years of repetitive loading with a high prevalence of injuries. 18 The first objective of this study was to assess the intra-rater and inter-rater reliability of the AHD measurement by US image analysis, comparing differences in reliability when using two different anatomical landmarks in neutral shoulder position. Secondly, the effect of the activity level of the shoulder muscles on the AHD measurement were assessed. Third, potentially confounding factors such as strength and ROM contributing to adaptive changes due to overhead motion and their relation to the AHD were examined in groups with different training experience: competitive amateur overhead athletes and non-overhead athletes (NOHA).

| Study design and participants
The study protocol was approved by the Ethics Committee of the Faculty week. The NOHA group (n = 13) included athletes (e.g., fitness, running and soccer), who did not perform any movements typically associated with overhead sports such as throwing or spiking movements (Table 1).
To identify shoulder impairments, all participants were asked to complete the German version of the Quick Disabilities of the Shoulder, Arm and Hand score. OHA were additionally asked to complete the German version of the Kerlan-Jobe orthopaedic clinic shoulder and elbow score (KJOC-G) 19 to identify impairments, activity limitations or participation restrictions, which was confirmed by KJOC-G scores of 98.6 ± 1.6 for the dominant and 98.4 ± 2.1 for the non-dominant arm.
The participants took part in one experimental session, in which anthropometric measures (body height, body mass, arm and forearm length), shoulder muscle strength, ROM and AHD were determined. Furthermore, the ratio of peak isometric external-to internal-rotation strength (ER/IR ratio) was calculated. Data of each movement direction were averaged over the three trials.

| ROM measurement
The active glenohumeral internal and external ROM of both shoulder joints were measured by an experienced physiotherapist   For inter-rater reliability the same images were additionally analysed by a blinded trained physiotherapist. The SE of measurement SEM = SD Â √ (1 À ICC) and the minimal detectable change MDC 95% ¼ 1:96 Â ffiffiffi 2 p Â SEM were calculated. 28 Group and sex differences were analysed using independent t-tests. Differences between the shoulder sides were analysed by

| RESULTS
Demographic characteristics did not differ significantly between OHA and NOHA (Table 1).

| Reliability of the AHD image analysis
The intra-rater reliability for repeated quantification of the AHD by the same rater was excellent (ICC 3,1 0.996; CI95% 0.991-0.998), with absolute mean differences being 0.02 ± 0.30 mm. The SEM was 0.02 mm and the MDC was 0.05 mm.
For both AHD quantification approaches, the inter-rater reliability was excellent (distance 1: ICC 2,1 0.997; CI95% 0.993-0.999; distance F I G U R E 1 (A) Ultrasound measurement setting for AHD measurement in neutral shoulder position; (B) Schematic illustration of the two different distances (1) the shortest distance between the most infero-lateral edge of the acromion and the most superior aspect of the humerus and (2) the perpendicular distance (90 ± 5 ) between the most infero-lateral edge of the acromion and the humeral head; (C) Ultrasound image for AHD measurement in neutral shoulder position. AHD, Acromiohumeral distance 2: ICC 2,1 0.959; CI95% 0.759-0.988). The absolute difference between the mean values obtained by the two raters was significantly smaller for distance 1 (0.02 ± 0.27 mm) compared to that for distance 2 (0.37 ± 0.42 mm) (p = 0.01) (Figure 2). The SEM and MDC for distance 1 were 0.02 mm and 0.04 mm respectively and those for distance 2 were 0.08 mm and 0.23 mm respectively.

| Effects of shoulder muscle activity level
The AHD absolute values and the AHD changes with abduction in any measurement position and shoulder side did not significantly differ between the OHA and the NOHA groups ( Figure 3). In both groups,

| Relationships of shoulder strength and ROM with AHD
There were no significant differences in absolute strength, normalized peak torque values and ratios or ROMs between OHA and NOHA ( Table 2).

| Sex differences
Regarding anthropometric measures, as expected, males were compared to females displaying a significantly larger body height

| DISCUSSION
This study focused on the methodological aspects of the AHD measurement obtained with ultrasonography.

| Reliability of the AHD image analysis
We tested two different ultrasound methods of AHD measurement that had been previously been described in the literature in enough detail to be replicated. 13,14 The agreement between the two different AHD measurement approaches justifies the application of both analysis techniques. However, based on the smaller absolute mean difference, using the shortest distance between the most infero-lateral edge of the acromion and the most superior aspect of the humerus as a landmark (distance 1) should be preferentially used when measuring the AHD on ultrasound images.
Our inter-rater reliability was similarly high as in a previous study. 30 However, high reliability may depend on short time intervals between the measurements and the degree of shoulder pathology.
Long time intervals between image analyses may reduce reliability.
For example, when images were analysed six months apart, inter-rater reliability (ICC 0.50) and intra-rater reliability (ICC 0.56 and 0.57) was just moderate. 11 Pathologies may similarly reduce reliability as bony structures used as distance borders might be more difficult to identify due to inflammatory reactions and soft tissue alterations. 11 That our study focused on healthy shoulders, which were analysed one week apart, may thus have contributed to higher reliability.

| Effects of shoulder muscle activity level
Our study demonstrated that increased shoulder muscle activity increased the AHD reduction with abduction. Active and loaded positions led to larger AHD reductions compared to passive positions. This is in accordance with Thompson

| Relationships of shoulder strength and ROM with AHD
Larger ROM may impair the preservation of the subacromial space and subsequent prevention of impingement-related conditions. Our study exclusively showed negative correlations between ROM and AHD. This is in agreement with Mackenzie, Herrington, 33 who showed that the AHD was preserved less with larger ER-ROM and TROM in non-athletes. As shoulder hypermobility is assumed to contribute to the development of a subacromial impingement, 7 enlarged shoulder ROM may be associated with increased risks of impingement-related conditions. As a consequence the detection of enlarged shoulder ROM warrants joint stabilizing interventions.

| Effect of training experience on AHD
The finding of no significant differences between the AHD in the OHA and NOHA groups suggests that loading-dependent adaptations in NOHA do not impact on the AHD. Other studies have reported greater AHD in college baseball players compared to controls, 32 and shorter AHD in tennis players compared to controls. 35 These inconsistent results across studies of overhead athletes suggest factors other than training experience are impacting on AHD measurements.

| Limitations
Reliability was tested only for the image analysis process of the AHD measurements; therefore results of this study cannot be extended to the image acquisition process of the measurements. We did not measure abduction angles higher than 60 , which may better reflect sportrelated movements, as acoustic shadows typically occur in higher ranges of shoulder abduction. 13 As a result, potential adaptations or changes may have been missed, if those occurred at higher degrees of abduction only. Scapular kinematics have not been measured within this study. However, scapular movement is considered to affect the AHD in a manner of adapted kinematics that may preserve the AHD for example, in athletes. 36 Furthermore, bony 37 and tendinous 32 characteristics and muscular modifications have not been taken into account. Anatomical abnormalities could thus have affected the subacromial space and its variation over position changes in individual cases. Moreover, this study investigated healthy shoulders only.
Results and assumptions of this study cannot be extended to pathological shoulders.

| CONCLUSIONS
Measuring the shortest distance between the most infero-lateral edge of the acromion and the most superior aspect of the humerus can be advised as preferred image analysis procedure for AHD measurement.