Study design
A retrospective review was conducted of prospectively collected data on patients who underwent TSA or RSA between January of 1992 and May of 2012 at the University of Florida. Both TSA and RSA were included in this analysis as RSA represents approximately one-third of the nationally reported shoulder replacement surgeries in the U.S. (Westermann et al. 2015). Diagnoses included osteoarthritis, acute fracture, rotator cuff tear, and rotator cuff arthropathy, and other (osteonecrosis, avascular necrosis, and synovial chondromatosis).
Of the enrolled patients, 310 had complete follow-up data for our research questions. Patients underwent TSA or RSA and had a minimum 2-year follow up and a final mid-term mean follow up of 5.0 ± 2.5 years (range 3–17 years). Age ranged from 21 to 95 years. Study documents and procedures were approved by the University of Florida Institutional Review Board and procedures conformed to the guidelines set by the Declaration of Helsinki for the treatment of human subjects.
Patients were stratified into groups based on BMI values: non-obese (BMI < 30 kg/m2), obese (30–39.9 kg/m2), and morbidly obese (≥40 kg/m2) for statistical analysis (Kane et al. 2010). Demographics (age, height, weight, BMI, and ethnicity/race), surgery type (TSA or RSA), comorbidities, perioperative analgesic use, and tobacco use were collected preoperatively.
Patient-reported outcome measures
Surveys were administered to each participant on subjective shoulder-specific functional and pain outcomes, perceived physical activity, and general health.
American Shoulder and Elbow Surgeons Evaluation form (ASES)
The ASES permits the patient to self-evaluate pain, shoulder instability, and activities of daily living (Richards et al. 1994). Eleven items are used to generate a score (pain has one item, function ten items). The physician-assessment section documents ROM, physical signs, strength, and stability. The final score is derived from the pain and cumulative activities of daily living score (Richards et al. 1994). A minimal clinically important difference is 6.4 points (Roy et al. 2009).
Shoulder Pain and Disability Index (SPADI)
The SPADI assesses shoulder pain and disability in the outpatient setting (Breckenridge and McAuley 2011), with 13 items categorized into two subscales: pain (five items) and disability (eight items) (Roach et al. 1991). This instrument has high internal consistency with Chronbach’s α values ranging from 0.86 to 0.95 for the subscales. The SPADI responds to change over time in various patient populations, and can discriminate between patients with improving and deteriorating conditions (Beaton and Richards 1996; Roy et al. 2009; Williams et al. 1995). A minimal clinically important difference is 8 points (Roy et al. 2009).
University of California at Los Angeles Shoulder Rating scale (UCLA)
This scale has separate domain areas: pain, function, strength of forward flexion, active forward flexion, and overall patient satisfaction (Amstutz et al. 1981). The scale is weighted by pain, and function accounts for 20 points. The other domains account for 15 points, giving a total score of 35 points (<27 is poor and fair and >27 is good to excellent). This has been used in patients with shoulder replacement due to osteoarthritis, rotator cuff disease, and instability (Kirkley et al. 2003).
Constant score
This score is comprised of subjective assessments relating to pain, activities of daily living, and objective measures of shoulder motion and strength. The maximum score is 100 points, with higher scores representing less pain and better shoulder function (Constant et al. 2008). Construct validity is 6 = .673, and the Constant score correlates moderately well with the UCLA score (r = 0.673) (Oh et al. 2009).
Medical Outcomes Short Form 12 (SF-12)
The SF-12 is a standardized, validated health related QOL questionnaire comprised of 12 items of the Medical Outcomes Short Form 36 original instrument (Ware et al. 1996). The SF-12 summary score (PCS-12 and MCS-12) ranges from 0 to 100 points. The SF-12 has a high test–retest reliability (r = 0.89) and has been validated for use in independently living older adults with chronic illnesses (Resnick and Nahm 2001).
Radiological outcomes and adverse events
Development of radiolucency lines during the follow-up (specifically within the glenoid) and scapular notching was identified by the surgeon in the study (Choi et al. 2013). Adverse events included infections, bone graft failure, loosening, dislocation, rotator cuff failure, wound complications, fracture, and pain and stiffness. The number of revisions in each surgery category was documented. The criteria for revisions included the following: infections, loosening or instability, dislocation, bone graft failure, and rotator cuff repair failure. If revisions occurred, the follow-up time was determined from the initial procedure.
Objective functional measures
Passive and active shoulder ROM was assessed by an experienced athletic trainer using a goniometer. Active external rotation and active elevation measures were performed by the patient. Each participant lifted progressively higher weights in increments of 2.2 kg during a lateral straight arm raise to estimate strength. The strength value was defined as the maximal weight the patient could lift with appropriate form once.
Statistical analyses
Data were analyzed using the Statistical Package for the Social Sciences (SPSS version 24.0, IBM Corp; Chicago, IL). Patient characteristics were compared between the three BMI strata (non-obese, overweight, and obese) using one-way analysis of variance (ANOVA) and a Scheffé post hoc analysis for continuous variables and by Kruskal–Wallis tests for categorical variables. Chi square tests were used to determine whether distributions of the shoulder pain mechanisms, radiolucency lines, and scapular notching were different among the groups. To address the purposes of the study, generalized linear models were used, where the dependent variables were study outcomes (ASES, SPADI, UCLA, Constant scores, SF-12, range of motion and strength) and the predictors were obesity status (non-obese, obese, and morbidly obese) and surgical type (RSA or TSA). Covariates were diagnosis and age. Significance was established at p < .05 a priori for all tests.