Procedure
A convenience sample of 102 patients with unilateral chronic anterior shoulder pain (more than three months), and with clinical symptoms of anterior shoulder pain, was recruited from three different primary care centers. General practitioners (GPs) carried out the recruitment. Then, research assistants assessed participants for eligibility. If participants satisfied the inclusion criteria, then they were studied. Five participants declined to participate, and 10 participants did not meet the inclusion criteria, hence, a sample comprised of 87 participants was assessed. Research assistants collected the informed consent for every participant.
All participants in the study gave their written informed consent. Participants had to meet the following inclusion criteria to be classified as anterior shoulder pain [9, 16, 17]: i) positive cross-arm test; ii) painful arc of movement during forward flexion and/or internal rotation; (iii) elicitation of tenderness throughout palpation of the coracoid process.
Furthermore, other inclusion criteria had to be met: both men and women aged between 18 and 55 years; no history of significant shoulder trauma, such as fracture or clinically/ultrasonographic-suspected full thickness rotator-cuff tear. Participants were excluded from this study if any of these conditions were presented: (i) recent shoulder dislocation, systemic illnesses such as rheumatoid arthritis, and evidence of adhesive capsulitis, as indicated by passive range of motion loss > 25% in 2 planes of shoulder motion, and loss > 50% in passive external rotation; (ii) shoulder pain that was deemed to be originating from any passive and/or neck movement or if there was a neurological impairment, osteoporosis, haemophilia and/or malignancies; (iii) shoulder surgery in the last year, (iv) corticoid injections during the 6 months prior to the study; (v) analgesic-antiinflamatory medication intake during 48 h prior to the study.
Outcome measures
Coracohumeral distance (CHD)
A diagnostic ultrasound unit, Sonosite M-turbo (GE Healthcare, Wauwatosa, WI) with a dynamic range up to 165 dB, was used. Furthermore, a 6–13-MHz linear transducer with 196 piezoelectric crystals with a specific ultrasound system called “SonoMB® multi-beam imaging”, to increase resolution and improve visualisation of physiological and subtle tissue differences, was used to capture images in a grey scale of 256 shades. Ultrasound images were obtained by a single examiner, who was a licensed physiotherapist with advanced training in musculoskeletal ultrasound imaging, and 4-years of experience. Three measurements were taken. An interval of 1 minute was provided between measures, encouraging the patient to move freely. Patients were then repositioned and the second and third set of measurements was successively taken. The ultrasound examiner was blind to all measurements (values were obscured by placing a sticker on the ultrasound screen, meanwhile a research assistant took them and put into a dataset). All the ultrasound measures were expressed in centimeters. CHD was measured at 0 and 60 degrees of active shoulder elevation in the scapular plane, neutral shoulder rotation, with the participant seated in an upright position.
Patients were seated upright without back support, their feet flat on the ground. To guarantee 0 and 60 degrees shoulder elevation, a hydro-goniometer was placed on the patient’s arm [18]. CHD was defined as the shortest linear distance between the coracoid and the adjacent humeral head [9]. The ultrasound transducer was placed over the most anterior aspect of the shoulder, observing the coracoid process and the humeral head on the screen, taking the shortest distance between them. CHD was measured in centimeters, using the calipers on the ultrasound screen (Figs. 1 and 2).
ROM-free of pain at shoulder elevation
Range of movement (ROM) free of pain at shoulder elevation was taken using the same procedure as followed for CHD ultrasonography measures, excepting a change in the patient’s position (stand up position). Three measures were taken separated by an interval of 1 minute, and mean was calculated. ROM was expressed in degrees.
Shoulder pain and disability index (SPADI)
The Shoulder Pain and Disability Index (SPADI) [19] is a self-administered questionnaire that consists of two dimensions, one for pain and another for functional activities. SPADI total score fluctuates between 0 and 100, with 0 = best and 100 = worst. SPADI has shown to have good internal consistency (overall Cronbach’s alpha = 0.95; for the pain subscale = 0.92; for the disability subscale = 0.93), as well as the ability to detect change over time [20].
Data analysis
The Statistical Package for the Social Sciences was used for analyzing the collected data (version 23.0 for Mac; SPSS Inc. Chicago, IL). Normality of the variables was visually tested for a Gaussian distribution and additionally tested with a 1-sample Kolmogorov- Smirnov goodness-of-fit test.
For the calculation of reliability of CHI the model or a 2-way mixed consistency intraclass correlation coefficient (ICC) model was used. Hereby a reliability coefficient less than 0.50 was an indication of “poor” reliability; “moderate” between 0.50 and 0.75, “good” between 0.76 and 0.90; and “excellent” over 0.90 [21]. The Standard Error of Measurement (SEM) and the minimal detectable change (MDC) with 95% confidence bounds (MDC95) were calculated. The MDC has been defined as the minimal amount of change that is required to distinguish a true performance change from a change due to variability in performance or measurement error [22]. To the best of knowledge, there is no studies reporting the MDC in the calculation of intra-rater reliability for CHD.
To determine the correlation between CHD at 0 and 60 degrees with SPADI, and ROM free of pain in scapular plane, Pearson correlation coefficient was calculated for normally distributed data, or Spearman’s coefficient in case of absence of normality. Strong correlation was defined as values greater than 0.7; between 0.5 and 0.7 correlation was considered moderate; between 0.3 and 0.5 was considered weak correlation [23].