Background: We sought to characterize humeral-sided radiographic changes at a minimum of 2 years after reverse
shoulder arthroplasty (RSA) to determine their association with specific implantation techniques.
Methods: The immediate and most recent postoperative anteroposterior radiographs of 120 shoulders with primary RSA
and a minimum of 2-years of radiographic follow-up were analyzed (mean follow-up 35.2 months). Stress shielding was
evaluated by measuring cortical thickness at 4 different locations. Three independent examiners evaluated radiographs
for humeral osteolysis, radiolucent lines, stress shielding, stem loosening, and scapular notching.
Results: The cortical diameter, marker of external stress shielding, significantly decreased from initial to most
recent measurement (P<0.001), but did not differ between cemented and uncemented groups. Cemented stems had
significantly more osteolysis and radiolucent lines; uncemented stems had significantly more internal stress shielding
(P<001). The presence of scapular notching was significantly correlated with the presence of humeral osteolysis
(P<0.001). Three (2.5%) stems were deemed “at risk” for loosening and 2 (1.7%) were loose.
Conclusion: Cemented humeral stems were associated with an increased rate of radiolucent lines and osteolysis,
whereas uncemented stems were associated with more internal stress shielding. Humeral cortical thickness significantly
decreased over time regardless of fixation. There was an association between scapular notching and increased
Level of evidence: III