Radiographic Severity May Not be Associated with Pain and Function in Glenohumeral Arthritis.

Abstract

Despite the routine use of plain radiographs to stratify the severity of glenohumeral osteoarthritis, little is known about the relationship between radiographic measures and patient-perceived pain and function. (1) What radiographic findings are associated with worse pain and function in patients with glenohumeral osteoarthritis? (2) What demographic factors are associated with worse pain and function in patients with glenohumeral osteoarthritis? This retrospective study included patients presenting for an initial office visit for primary glenohumeral osteoarthritis. Patients with other concurrent shoulder pathologic findings, prior surgery, lack of pain and functional scores, recent injection, or inadequate radiographs were excluded. Between January 2017 and January 2019, 3133 patients were eligible based on these inclusion criteria; 59% (1860) had outcome assessments and 48% (893) of those had radiographs. An additional 42% (378) of those with radiographs were excluded because of other shoulder findings, recent injection, prior surgery, or inadequate radiographs, leaving 16% (515 of 3133) who were fully analyzed in this study. A radiographic review included the joint space width, posterior humeral head subluxation, inferior humeral head osteophyte size, cystic change, and head asphericity. Additionally, radiographic arthritis was classified according to the Walch, Samilson-Prieto, and Kellgren-Lawrence classifications by two separate reviewers. Radiographic and demographic criteria as well as the presence of psychologic or mental illness were correlated with VAS Pain (range 1-10; minimal clinically important difference [MCID] 1.6), American Shoulder and Elbow Surgeons (ASES; range 0-100; MCID 13.6), Single Assessment Numeric Evaluation (SANE; range 0-100; MCID 14), and Simple Shoulder Test (SST; range 0-12; MCID 1.5) scores using univariate and multivariable regression analyses. After accounting for age, gender, and psychologic illness in the multivariable analysis, we found that patients with Samilson-Prieto Grade 4 arthrosis had lower VAS Pain scores (β = -1.9; p = 0.02) than those with Grade 0 or 1 did; however, no clinically important associations were found between Samilson-Prieto Grade 4 and ASES (β = 7; p = 0.25), SANE (β = 4; p = 0.63), or SST (β = 0.5; p = 0.62) scores. No clinically important associations were found between Kellgren-Lawrence Grade 3 and VAS Pain (β = 1.4; p = 0.10), ASES (β = -8; p = 0.22), SANE (β = -13; p = 0.11), or SST scores (β = 0.4; p = 0.66). Radiographic joint space and posterior subluxation also did not have any clinically important associations with VAS Pain or functional scores. In assessing Walch glenoid type, there was no clinically important association between glenoid type and VAS Pain (F = 3.1; p < 0.01), ASES (F = 1.9; p = 0.15), SANE (F = 0.45; p = 0.66), or SST scores (F = 0.76; p = 0.71). Men had higher SST scores than women did (β = 2.0; p < 0.01), but there were no clinically important differences in VAS Pain (β = -0.4; p = 0.04), ASES (β = 6; p < 0.01), or SANE (β = 4; p = 0.07) scores. No clinically important association was found between age or the presence of any psychologic illness and VAS Pain or functional scores. In patients with glenohumeral arthritis, no consistent clinically important differences in pain or function were discovered with respect to radiographic or demographic factors. Surgeons should understand that the pain levels of patients with glenohumeral arthritis may not parallel radiographic severity. Future studies can build on these findings by examining other non-radiographic or demographic factors that affect pain in patients with shoulder arthritis, such as psychological factors. Level III, prognostic study.